Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Design Criteria of Soft Exogloves for Hand Rehabilitation-Assistance Tasks

Design Criteria of Soft Exogloves for Hand Rehabilitation-Assistance Tasks Hindawi Applied Bionics and Biomechanics Volume 2020, Article ID 2724783, 19 pages https://doi.org/10.1155/2020/2724783 Review Article Design Criteria of Soft Exogloves for Hand Rehabilitation- Assistance Tasks 1 1 1 Juana-Mariel Dávila-Vilchis , Juan C. Ávila-Vilchis , Adriana H. Vilchis-González , 1,2 and LAZ-Avilés Faculty of Engineering, Universidad Autónoma del Estado de México, Toluca 50130, Mexico Cátedras CONACYT, Universidad Autónoma del Estado de México, Toluca 50130, Mexico Correspondence should be addressed to Adriana H. Vilchis-González; avilchisg@uaemex.mx Received 24 February 2020; Revised 3 July 2020; Accepted 7 July 2020; Published 1 August 2020 Academic Editor: Simo Saarakkala Copyright © 2020 Juana-Mariel Dávila-Vilchis et al. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. This paper establishes design criteria for soft exogloves (SEG) to be used as rehabilitation or assistance devices. This research consists in identifying, selecting, and grouping SEG features based on the analysis of 91 systems that have been proposed during the last decade. Thus, function, mobility, and usability criteria are defined and explicitly discussed to highlight SEG design guidelines. Additionally, this study provides a detailed description of each system that was analysed including application, functional task, palm design, actuation type, assistance mode, degrees of freedom (DOF), target fingers, motions, material, weight, force, pressure (only for fluids), control strategy, and assessment. Such characteristics have been reported according to specific design methodologies and operating principles. Technological trends are contemplated in this contribution with emphasis on SEG design opportunity areas. In this review, suggestions, limitations, and implications are also discussed in order to enhance future SEG developments aimed at stroke survivors or people with hand disabilities. because cartilage weakens, muscle mass decreases, and joint 1. Introduction stiffness increases [5]. More than 50 million elderly people Hand and finger motions are imperative for grasping and have difficulties to achieve accurate gripping and pinching manipulation tasks. Nonetheless, people who have suffered forces, and their range of motion (ROM) is limited as well from cerebral palsy (CP), stroke, or spinal cord injury (SCI) as their work area [6]. Therefore, people with hand disabilities can initiate a have great difficulty in accomplishing these activities of daily living (ADL) by themselves. A person with any of these prompt rehabilitation protocol in order to start recovering pathologies could present clenched fist, spasticity, uncoordi- motor skills, stop joint stiffness, and increase their indepen- nated motions, loss of strength, or diminished dexterity. dence and self-esteem [7]. Physical and occupational thera- These are consequences of a neuronal impairment that is pies are the most common treatments to recover patients’ responsible for controlling motricity, muscle endurance, movements, for example, adduction-abduction or flexion- and tonicity [1]. Worldwide, more than 15 million people extension of finger, wrist, or elbow joints. However, these are affected each year [2], and only 11.6% of the stroke survi- routines can be exhausting, time-consuming, and, relatively, vors are able to recover dexterity [3]. Patients with these dis- costly since patients require the assistance of a therapist abilities can, freely, flex their hand muscles but show whose availability is uncertain [8]. abnormal resistance when extending them [4], requiring Normally, rehabilitation programs are customized for physical rehabilitation or assistance. each patient due to their impairment, age, and anthropomet- Other hand motor deficits are caused by ageing or hand ric dimensions. Moreover, these robot procedures are classi- deformities such as rheumatoid arthritis or osteoarthritis, fied into three main assistance levels: passive assisted mode 2 Applied Bionics and Biomechanics SEG design criteria Usability (ix) Modularity Mobility (iii) Actuation (iv) Materials (x) Portability Function (xi) Customization (v) Motion (vi) Manufacture guidance (xii) Mode of (i) Rehabilitation intervention (vii) Operation (viii) Assessment & control (xiii) Costs (ii) Assistance Figure 1: Classification of soft exo-gloves design criteria. guidelines for SEG developments based on an extensive (PAM), active assisted mode (AAM), and active resistive mode (ARM) depending on the recovery status of patients review of the state of the art and of the technique from the and support of a robot [9]. last decade. Moreover, a detailed description of 91 SEG sys- Literature has reported that rehabilitation protocols can tems is provided along with implications, limitations, and be executed by robots or soft wearable devices which have suggestions for future developments. This paper is organized as follows. Section 2 presents, emerged as a therapy tool with safe human interactions, low weight, and affordable systems [10]. Particularly, SEG classifies, and discusses the criteria that are proposed for have become an alternative approach in the effort to over- SEG design based on reported devices and specific literature. come hand dysfunctions and assist patients with handling Section 3 reports SEG’s development guidelines together tasks. SEG have the ability to combine conventional therapy with the characteristics of the 91 reviewed devices. Section 4 provides a discussion concerning significant aspects (limi- with wearable systems to mimic the natural movement of fin- gers in order to increase their mobility, preventing spasticity tations, implications, and suggestions) to be taken into and joint stiffness [11]. account for future developments of SEG systems. Conclu- SEG have mainly evolved in terms of their design, fabri- sions are at the end of this document in Section 5. cation, and control [12]. Pioneering designs started using sport gloves incorporating a control system [13, 14]. Then, 2. SEG Design Criteria SEG proposals explored synthetic leather [15], rubber [16, 17], and fabrics [18, 19] to provide flexible human-robotic Hand mobility characterization in SEG designs has turned interactions as in the case of bike gloves [20]. Elastomers out to be a challenge since hand anatomy is one of the most have become the primary option to empower flexibility and complex kinematics parts of the human body with 20 DOF lightness [21]. Moreover, instead of closed palm designs for the whole of the fingers: one for abduction-adduction in (CPD) where the whole hand is covered with the glove, open every finger (thumb included); 12 for flexion-extension for palm designs (OPD) with bare hands use elastomers trying to index, middle, ring, and pinkie fingers; and three for thumb behave as a natural extension of the human hand to compete including opponent motion [29]. with skin properties in order to achieve a suitable contact In this paper, 2 function criteria, 6 mobility criteria, and 5 with objects [22, 23]. Other assistance SEG have been devel- usability criteria are proposed in order to enhance SEG oped for material handling in hazardous environments, sup- designs and enable fast developments. These design criteria port in heavy-lifting tasks [24, 25], or extravehicular tasks in are based on the aspects that have been identified from the space [26]. 91 SEG systems reported in this article and on the soft wear- Mostly, SEG systems have been driven by electrical able device’s methodology established in [28]. Figure 1 illus- energy or fluid (pneumatic or hydraulic) pressurization. trates the proposed criterion classification. Regarding electrical power supplies, tendon-driven systems Moreover, biological inspiration has come to the fore in employ linear actuators to push and pull cables embedded SEG design to emulate an animal’s motion looking for stabil- in Teflon tubes [27]. Pneumatic actuation includes fiber- ity [30] or optimal grasping tasks [31]. According to [8], SEG reinforced elastomer actuators (FREAs), inflatable chambers, should weigh less than 500 g, provide easy and comfortable or pneumatic artificial muscles, commonly known as McKib- donn-doff, and achieve 10 open-close finger cycles per min- ben muscles [28]. ute for effective actuation. Regarding SEG mechanical design, People with hand dysfunctions demand for reliable SEG authors in [29] suggest taking into account the number of to improve their quality of life. Nevertheless, the lack of joints and working DOF, the type of actuators, and the appli- affordable and accessible SEG for hand impairment patients cation. Other attributes in SEG design should adopt the char- with low-cost manufacturing processes is still a significant acteristics of a rehabilitation device which include mode of challenge. Therefore, this paper has reviewed the progress intervention (unilateral or bilateral), number of DOF, target in the field of SEG for neuromuscular rehabilitation and portion (distal, proximal, or quantity), and motion guidance assistance to overcome hand motor dysfunctions. (passive or active), among others [32]. The main contribution of this paper is the identification Based on reported literature, the next paragraphs discuss and classification of 13 design criteria to provide a set of the criteria presented in Figure 1. Applied Bionics and Biomechanics 3 SEG for assistance tasks are designed to perform three 2.1. SEG Function Criteria. SEG are classified into rehabilita- tive or assistive devices depending on their purpose [14]. SEG integral functions of the human hand: (i) finger mobilization, systems must be able to execute physical therapy and manip- (ii) holding (grasping and gripping) with high precision and strength, and (iii) manipulation for positioning and releasing ulation tasks to offer efficient and competitive devices for those with hand disabilities. Then, rehabilitation and assis- objects [8]. Assistive SEG should execute grasping, holding- tance criteria must consider the aspects discussed in the lifting, and releasing motions as continuous actions to respective paragraphs. achieve a complete manipulation [50]. To achieve stable grasping, thumb, index, and middle fingers must be included on SEG systems [35]. According to [51], soft exoglove devices 2.1.1. Rehabilitation Criterion. Rehabilitation SEG are should provide 8 N of grasping force to manipulate an object designed to help the patient regain strength, dexterity, and with a mass of 1.5 kg. coordination to recover hand functionality and range of motion (ROM) [33]. These SEG are focused on performing 2.2. SEG Mobility Criteria. From a functional perspective, specific fist motions such as full, hook, straight, and tabletop authors in [52] propose that weight, size, and power con- [18] or open-close to improve grasping tasks [34]. sumption can define an efficient soft exoglove that fits the Thumb, index, and middle finger flexion-extension is anatomical ROM of the human hand. The mass of the whole needed for strong grasping [31, 35]. Supplementary motions system should not exceed 3 kg to be considered as an assistive such as adduction-abduction are required to grasp and device [50]. These characteristics are included in criteria 3 to release objects in a more natural way [36]. Furthermore, flex- 8 (see Figure 1): actuation, materials, guidance mode, manu- ion at the interphalangeal (IP) and metacarpophalangeal facture, operation and control, and assessment that are dis- (MCP) joints with rotation at the carpometacarpal (CMC) cussed as follows. joint is necessary to reproduce thumb opposition [37]. Other SEG are able to perform wrist flexion [38], wrist radial-ulnar 2.2.1. Actuation Criterion. As aforementioned, tendon- deviation [16], or forearm pronation-supination motion [39]. driven actuators use wires to emulate human tendon func- SEG rehabilitation routines can include virtual reality in tions as flexion-extension motion. This type of actuation order to analyse the effects of brain stimulation when execut- can include Bowden cable transmissions to separate the con- ing specific tasks [40]. Patients are immersed in a game envi- trol unit from the end effector and reduce weight [53]. Also, ronment where they achieve manipulation tasks such as artificial muscle wires have been proposed to avoid friction squeezing oranges, catching butterflies, or grabbing objects [54], and shape memory alloys (SMA) have been employed [39]. Other SEG rely on neuroimaging techniques [41] or due to their elasticity [55] and high force-weight ratio [26]. provide feedback to assess a patient’s conditions and monitor On the other hand, pneumatic actuators could be embed- their progress [42]. Nevertheless, it is not enough to train the ded into inflatable air bladders [16] and into a double layer brain and do physical therapy; a successful rehabilitation sheet with curved rubber muscles [15] or made of flexible process depends on the patient’s response and their own electrostatic discharge plastic sheet materials [1, 56]. The capabilities [43]. McKibben muscles represent an affordable choice [57] and Depending on each rehabilitation protocol, the required have the ability to constrain any radial expansion during time to use a soft exoglove varies. For instance, 60 minutes pressurization [58]. Hydraulic actuators offer high load per day is recommended by [44]. Pilot tests performed by capacity [11]. [45] suggested rehabilitation sessions from 30 to 40 minutes A new trend is hybrid actuation which fits hand motion 5 days a week. Authors in [38] recommend 45 minutes but shape using soft pneumatic actuators and tendon-driven no more than 90 minutes per day to avoid SEG strain defor- operation [7], providing customization based on rigid frames mations. Authors in [46] suggest 180 minutes per week, while and soft muscles [48]. Table 1 reports the advantages and dis- authors in [39] determine that 30 minutes per day over the advantages of different SEG actuations. course of 20 sessions is necessary for a positive sizable impact When using a soft glove, patient safety must be guaran- on the impaired hand. Furthermore, to achieve a successful teed. Thus, all SEG must include different safety strategies rehabilitation program, patients should combine 30 minutes and levels in their design. For example, on cable actuation, of SEG training with 30 minutes on occupational therapy per mechanical stops, torque, or tension limiters have been day [47]. implemented [59]. Regarding pneumatic actuation, solenoid and exhaust valves are employed along with pressure regula- 2.1.2. Assistance Criterion. Eating, dressing, and writing are tors to control air flow or avoid air returns [41]. Quasistatic, everyday actions that are done unconsciously. Nevertheless, dynamic, and material failures are discussed in [60], where those tasks turn out to be a tough challenge for people with measures that can be considered in order to avoid unsafe sit- hand dysfunctions. Normally, patients depend on their fam- uations for soft robots are provided. ily or on a therapist to assist them [48]. Hence, assistive SEG Other safety levels have been applied to the electrical con- are intended to help patients to achieve manipulation tasks figuration such as emergency stops, watch dogs, or physical despite their restricted ROM, to interact with their surround- decoupling of power interfaces from logic ones by electromag- ings, and to execute ADL by themselves. These systems are netic couplings [51]. In addition, by using closed-loop control recommended when rehabilitative SEG are not enough to (CLC) schemes, sensing errors are minimized and operation overcome patient stiffness [49]. in a stable regime is ensured to avoid hyperextension at the 4 Applied Bionics and Biomechanics Table 1: Actuators for SEG systems. Actuation Types Advantages Disadvantages (i) Cable paths reduce friction (i) Complex transmissions (i) Muscle wires (ii) Provides continuous force (ii) Continuous hours of operation are restricted Electrical (ii) Tendon-driven (iii) Stores energy (iii) Nonlinearity of the system makes control (iii) Shape memory alloys (iv) Commercial availability difficult (i) Requires compressed air (i) FREA (i) Allows multiple DOF (ii) Requires a reservoir (ii) Supports their structural shape (ii) Inflatable chambers Pneumatic (iii) Inaccurate forces (iii) Pneumatic artificial (iii) Allows adaptability (iv) Problems with leaks muscles (iv) Lightweight (v) Portability is restricted (i) High load capacity and power (i) Heavy systems supply (ii) Problems with leaks Hydraulic (i) Fluid chambers (ii) Low cost (iii) Portability is restricted (iii) Allows multiple DOF (iv) Requires a reservoir and a pump wrist or overflexed fingers, for instance [20]. At the program- grip strength and reduce pressure on the skin [51, 71]. Addi- ming level, haptic feedback is also included to prevent acci- tionally, silicon materials offer stable fastening and prevent dents [61]. slippage [72]. These synthetic polymers are easy to wash More specialized safety strategies related to robots can be and do not absorb sweat compared to textile materials [23]. considered, such as safety standards or means to guarantee SEG made of fabrics have low cost and offer minimal mechanical impedance to finger motion [73]. Hence, coated system dependability [62] as fault prevention, fault removal, fault forecasting, and fault tolerance [63]. Being safety a pri- fabric SEG systems with thermoplastic polyurethane (TPU) ority aspect, it constitutes a current research area by itself and actuators are recommended for customization and to avoid must be taken into account in the development of SEG sys- slipping or muscle expansion problems [74]. tems. Concerning rehabilitation robots, ISO-IEC 80601-2- Actuators made of fabrics work at lower pressures than elastomer actuators due to their inherent stiffness [75]. 78 must be taken into account. Many specialized documents are recommended for readers interested in this topic and for Therefore, several researchers have work on design, charac- researchers and engineers working in SEG design (see, for terization, manufacture, and evaluation of soft elastomer instance, [64–66]). actuators for hand [76–78] and wrist [79] rehabilitation. Additionally, relevant features for actuators have been To match and support finger flexion-extension, some designs include multisegment elastomers with fiber rein- identified in SEG literature or proposed in this paper. For instance, current developments have focused on improving forcement [80, 81] or corrugated fabric layers [41, 43] which actuator design to tackle more DOF [67]. During SEG assem- are pressurized from 70 kPa up to 375 kPa [75]. Other bly, the actuators are mounted into the dorsal side of the designs include rigid plastic hoops [67] or nylon strings hand to avoid finger movement obstruction [68] and can be [82] to avoid radial deformations in FREA. Material selection has also played a significant role in fas- removed from the glove [69]. Actuators must not affect the active ROM of the finger joints and should allow free motions tening the actuators to the glove or fingers in a safe way. with more contact area for grasping tasks in a compliant Mostly, SEG proposals have employed magnets [83] or straps manner [21]. made of Velcro® [8, 18], fabrics [84], and rubber [24]. Other Furthermore, actuators should take less than 4 s for full designs had opted for sewing the components [71] or sepa- rating the system from the actuators to reduce weight. Actu- grasping [1]. The length of actuators should not be longer than the length of the fingers to avoid mismatching problems ators can be attached to the wrist through elastomer bracelets between them [23]. Actuators with low power consumption [39] or synthetic hide covers [25, 31]. and continuous hours of operation are recommended. 2.2.3. Manufacture Criterion. Mobility is also determined 2.2.2. Material Criterion. To enhance SEG operation, by manufacturing processes since specific elements can researchers continue to seek compliant, flexible, and light- be obtained by particular methods that, additionally, can weight materials to easily conform hand-finger anatomy with determine the weight and dexterity of the system. Conven- the shape of an object [41]. Hence, the payload capacity of tional manufacturing procedures involve polymer casting molds [85], reinforcements and inclusions [11], additive elastomers has been exploited to obtain an elastic modulus similar to that of human tissues and avoid cumbersome manufacturing, thin-film manufacturing, shape deposition designs [70]. manufacturing, and bonding [86]. Nonferromagnetic materials such as nylon, neoprene, Mostly, 3D printing two-part mold has been employed for polyester, or synthetic leather have been selected as compli- SEG spacers [23], cable guides [73], and elastomer actuators [87] where one mold is used to create a fluid chamber inside ant and affordable options to increase conformability and Applied Bionics and Biomechanics 5 Assistance Control mode DOF Number of motion Palm Fingers design Tasks Actuation Application 40 Material Assessment Function Force Pressure Weight Figure 2: Frequency of SEG aspects reported in Table 3. the actuators and the second one is addressed to create a fabric adduction-abduction [85] and perform flexion-extension layer on top of the actuators [41, 43]. Nevertheless, low repeat- motions [39, 71]. More sophisticated SEG systems have already begun an age that allows patients to perform a ability is the main drawback during this process [48]. Recent developments involve thermomethods [34], desired movement. When patients are able to achieve func- inverse flow injection [42, 82], lost wax molding [88], or tional ROM, the system will have no effect on the hand fused deposition modeling with 3D printing at home to [41] or will create an opposite force to improve the power reduce SEG costs and facilitate its acquisition [89]. However, of the patient. there is still room to improve SEG materials and fabrication Most of the reported SEG systems focus on PAM, a few with low costs. on AAM as well as on the combination of active and passive New trends are oriented to hybrid designs where they modes (see Figure 2). combine rigid and soft components to obtain more hand poses and more DOF [90] and provide active training that 2.2.5. Operation and Control Criterion. SEG operation is encourages user participation [91]. defined by their type of actuation and their components. Tendon-driven wires require servomotors, gearboxes, spools, 2.2.4. Motion Guidance Criterion. SEG are designed to follow and force/torque sensors to move them. Pneumatic systems specific trajectories defined by a therapist depending on the require a compressor, electrovalves or proportional valves, impairment of the patient. These trajectories seek to achieve pressure sensors, or regulators. All these components are a functional ROM during both active and passive modes. controlled on a data acquisition board which is plugged to a SEG is aimed at promoting active finger flexion and pas- PC or uses Bluetooth as a communication interface for the sive extension to increase patient autonomy during eating or SEG system [74]. drinking tasks [46, 92]. In the active assistance mode, Different schemes have been proposed to operate and patients attempt to move their hand and SEG are an addi- control SEG systems; for example, in [14], Faulhaber 1226 tional aid to complete the desired ROM [93] whereas in the 006B motors, CompactRIO board, and LabVIEW® are used. passive assistance mode, the exoglove provides all the assis- Authors in [21] use DCX22 motors, a control board tance to guide the desired movement [94]. In the patients’ TMS320F2808®, and Simulink®. Additionally, graphical user force recovery processes, effective SEG systems should, actively, participate with intensive training based on active interfaces (GUI)® have been implemented as a communica- tion channel for SEG systems [88, 91]. A broad range of and repetitive practical motions [95]. operation and control possibilities exists to select microcon- SEG should combine active and passive mobilizations for successful hand rehabilitation. For example, authors in [15] trollers and interfaces relying on desired real-time response, accuracy, number of components involved in the operation provide active extension on each finger. In [8], SEG also exert passive extension with active flexion and thumb opposition and control strategies, and specific requirements of each for grasping tasks. Other systems include active finger SEG system. Number of soft exogloves RT AT AT-RT S troke Hand disabilities Other Grasping Pinching Both Closed Open Tendon-driven Pneumatic Hydraulic Passive Active Active-passive < 10 > 10 All Ind-Mid-Thu All except Thu Elastomer Fabrics Other < 100 g 100-200 g > 200 g < 10 N 10-20 N > 20 N < 100 kPa 100-300 kPa > 300 kPa CLC OLC Forces ROM Other 6 Applied Bionics and Biomechanics to move specific hand joints that are connected to the afore- Normally, open-loop control (OLC) and closed-loop control strategies are implemented during SEG operation. mentioned muscles. Due to stable behaviours, force myogra- OLC schemes have used springs [1, 34] or mechanical phy (FMG) signals have been proposed to control the intention of the movement on SEG systems [20]. switches [54] for manual operation where patients are able to drive an actuator to accomplish a specific task [96]. OLC Motor impairment scales are applied to evaluate patient strategies require the system to be stable by construction. ROM to determine SEG operation ranges before starting an To regulate the desired variables or to track specific trajecto- aided rehabilitation process. These scales serve for the eval- ries that ensure patient safety while using a soft exoglove, uation of the damage that each patient has. According to [38, 105], patients with an Ashworth spasticity index (ASI) CLC strategies are implemented [97]. To achieve acceptable motions in CLC schemes, sensors are directly attached to value less than or equal to three can use a soft exoglove. A SEG [98, 99] without the patient worrying about making modified Ashworth scale (MAS) value less than or equal to accurate movements. two defined the use of a soft exoglove for active flexion- Proportional (P) [68], proportional-derivative (PD) [100, extension, according to [106]. SEG operation is also based on the functional independence measure (FIM) of the 101], or proportional-integral-derivative (PID) [15, 71] con- trollers are widely implemented for flow and force regulation. patient by which the value goes from 1 to 7 depending on Pulse width modulation (PWM) signals have been used to the assistance intensity [45]. Thus, for values above 3, open and close solenoid valves [51] and can be implemented patients present more autonomy [36]. in many control strategies for different applications. Other 2.2.6. Assessment Criterion. To ensure patient safety and SEG kinds of controllers can be used depending on the system nature and on the task objective. For instance, nonlinear con- operation, several tools such as joint contractions [31, 54], trollers, fuzzy approaches, or optimal linear control schemes bending angles [71], 3D visual motion analysis [11], or opti- could be developed for specific SEG systems. For instance, cal ROM at specific joints [44] have been employed to evalu- [102] provides an interesting review of soft robotic manipu- ate SEG performance. Other methods have opted for using mathe lator control strategies that could be considered to be applied matical models together with the finite element in SEG systems. method (FEM) for hand and finger trajectory characteriza- SEG operation is based on force and position require- tion [67]. To assess patient satisfaction when using SEG sys- ments to emulate human hand functions. These require- tems, questionnaires have been considered [88]. ments, among others, are taken into account to define the SEG assessment can be also done based on the blocked control strategy to be synthesized. For example, SEG should [48], grasping [21], pinching [44], or fingertip [14] forces that have less than 10 minutes of setting time to become a useful are quantified using bottles, cups, balls, telephones, cans, or tool for therapists [103]. Regarding fluid actuation, 10 N to fruits with variable mass, size, and texture [44, 51]. For cylin- 15 N are required for grasping tasks [11, 41]. SEG must be drical objects, the diameters go from 50 mm to 120 mm [21, able to generate 7 N per finger or around 25 N on the whole 75] with a mass of 300 g [107]. Experimental tests on SEG hand with distributed forces along the fingers to minimize assessment have been carried out with dummy hands [71] pressure location points, according to [34]. Normally, actua- and healthy individuals [44] or combining healthy people tors with variable stiffness require 120 kPa for pinching and and stroke survivors [75]. Other SEG evaluations perform 160 kPa for grasping [18] while multisegment actuators tasks with/without a soft exoglove and compare them [46, require between 345 kPa and 400 kPa for flexion motion 92]. ROM data have been collected when using a soft exo- [51]. Desired joint ROM define positions to be reached by glove and without it [31]. the patient when using a SEG system and provide reference To assess hand function and ROM using SEG systems, variables to be controlled. patients undergo coordination and dexterity tests. For To evaluate SEG effectiveness in rehabilitation or assis- example, the Kapandji score is used to evaluate thumb tance tasks, surface electromyography (EMG) has been performance on pinching and grasping tasks [108]. SEG implemented to detect user movement intentions [53], point assessment also considers the motricity index test (MIT) out muscle contractions [16], control finger motion, and [105], the Fugl-Meyer assessment (FMA) [46], the nine- force level activation of muscles [90] since this is a noninva- hole peg test (NHPT) [38], the Jebsen-Taylor hand test sive procedure that prevents muscle injuries. (JTT) [44], the box and block test (BBT) [11], the Purdue During gripping tasks for finger flexion-extension, EMG pegboard test (PPT) [45], or some writing tasks [109]. signals are captured from the extensor digitorum communis For each patient, one or more of the aforementioned (EDC) muscle together with the flexor digitorum superficialis methods could be chosen by his/her motor impairment or (FDS) [20, 50] or with the flexor digitorum profundis (FDP) by the therapist in charge of the respective rehabilitation pro- muscle [43, 73] since these muscles have been used and tested tocol in order to assess SEG systems. to work properly when implementing EMG procedures and Some authors have focused more on statistical analysis due to the number of fingers they are connected with. Then, about user condition than SEG performance [5, 40]. They data obtained from a set of electrodes are amplified, filtered, seek for a specific target group, rehabilitation time, training quantified, and converted from analog to digital signals dur- tools, age, or gender, for instance. ing SEG use [104]. This electrical stimulation should be mon- itored at least every 10 minutes to avoid muscle fatigue [103]. 2.3. SEG Usability Criteria. To guarantee a friendly and com- EMG signals can be used as control inputs when it is required fortable SEG use, modularity, portability, customization, Applied Bionics and Biomechanics 7 more beneficial than unilateral mode since patients can inte- mode of intervention, and cost criteria must be considered to develop a soft exoglove with particular characteristics as easy grate healthy and paretic hand motions during rehabilitation to put on and operate, working in an intuitive way, and hav- therapy [75]. The bilateral mode is supported by a master- slave therapy concept where healthy limbs act as masters ing low cost. These criteria are discussed below. and soft devices as slaves [101]. Then, healthy limbs become 2.3.1. Modularity Criterion. SEG designs have opted for mod- a support for paretic limbs whereas devices working in the ular configurations to ease donn and doff as in the cases of unilateral mode only exercise the impaired limb [111]. Bilat- [21, 72, 83]. Connections can be assembled to work on tar- eral mode rehabilitation could be recommended by the ther- geted tasks, and actuators are mounted one by one [39]. apist as a function of the impairment. Then, SEG design Besides, modular designs for bending motions with deploy- could consider the mode of intervention depending on the able mechanisms have been adopted to reduce weight and associated rehabilitation protocol. allow natural motion [48]. SEG quality can be improved by a modularized system with relatively low cost customization, 2.3.5. Cost Criterion. It has been noted that researches are easy maintenance, and low power consumption [23]. Addi- more interested in the functionality of their products than tionally, modular architectures allow for the replacement of in their price, since only few works report SEG costs. How- feasible SEG components [89]. Based on this information, ever, SEG cost will determine one of the aspects for the suc- modularity is highly recommended as one of the main char- cess of an exoglove as a commercial product. Therefore, acteristics of SEG systems. designers could generate low-cost readily available SEG sys- tems. For instance, authors in [34] propose that the assembly 2.3.2. Portability Criterion. To cope with patients’ demands should cost less than $30 USD in order to be a competitive and to guarantee continuous rehabilitation protocols, the choice. Another proposal establishes that manufactuing and use of SEG outside clinics has become a main design con- electronics should be less than $200 USD [100]. According cern to foster external rehabilitation [38, 110]. Nevertheless, to [52], soft exosuits for the upper limb should cost less than to achieve this objective, SEG performance depends on the $1000 USD, $465 USD for the elbow, and $470 USD for the number of hours they can operate continuously without hand. A detailed description about the component cost of having a fixed power supply. According to [51], an effective these configurations could be found in [59]. SEG costs can soft exoglove should achieve, without problems, 2 hours of vary due to the type of actuation, the type of components continuous operation or from 4 to 6 hours of intermittent and materials, the weight, and the country where they were operation. developed [112]. Moreover, the runtime of batteries should be more than Remarkable results about cost analysis between conven- one hour in order to guarantee the development of a rehabil- tional and aided therapy show that SEG rehabilitation is itation protocol session [100] until its completion or exert more affordable than therapist assistance since the reported from 15 to 20 minutes of passive guidance [88]. Normally, cost associated with aided therapy is almost three times less lithium-polymer batteries are used since they can last 3.8 expensive than the conventional one [45]. hours of continuous operation [23, 51]. Currently, Neofect™, Glohera™, and Bioservo™ compa- Patients should take physical therapy sessions at reha- nies have already patented their systems which have been bilitation facilities as well as at home [46, 92] in order to commercially exploited for hand rehabilitation and assis- perform exercises on their own and not only depend on tance. However, these commercial systems are available only the availability of therapists [34]. SEG must be lightweight in some countries and are, relatively, expensive. Importation to allow their transportation [31, 73]. Thus, control unit and shipping costs must be added to final prices for countries boxes should be set up independently of the glove to min- and locations where these systems are not available. imize additional load [74]. Some power supply designs include waist belts [43, 51, 89], backpacks [73], boxes 3. SEG Design Guidelines [11, 50], vests [84], waist pockets [53, 59], pockets [44], or a separate section located on another part of the human Based on the information provided in Section 2, Table 2 sum- body [25, 31]. marizes some of the main aspects related to the 13 proposed design criteria for SEG developments. 2.3.3. Customization Criterion. As established by [18], con- At present, SEG approaches are focused on improving formability, adaptability, and customization are some fea- functionality, strength, DOF, and ROM for object manip- tures that can be taken into account to fit, properly, the ulation. Figure 2 and Table 3 provide information for each hand of a patient and generate a compliant soft exoglove. of the 91 SEG systems reviewed in this paper, associated Particularly, customization affects SEG operation since each with the following 15 aspects: (1) function: robot rehabili- finger length varies due to sex, age, and finger palm size tation (RT) or assistance tasks (AT); (2) application: hand [17]. Thus, fasteners [71] and Velcro® straps [8, 110] have disability, stroke survivors, or SCI; (3) task: grasping, been used to attach, conveniently, SEG to hands. Otherwise, pinching, or manipulation; (4) palm design: OPD or deviations from a nonappropriate size or form may restrict CPD; (5) type and number of actuators: tendon-driven, hand movement or cause discomfort during SEG use [21]. pneumatic, or hydraulic; (6) assistance mode: PAM or 2.3.4. Mode of Intervention Criterion. To increase hand func- AAM; (7) DOF per finger; (8) targeted fingers; (9) motions: tion rehabilitation, a bilateral mode in SEG systems results flexion-extension, adduction-abduction, opponent, ulnar/ 8 Applied Bionics and Biomechanics (6) AAM should be the priority motion guidance for Table 2: Proposed criteria and considerations for SEG system design. SEG rehabilitation Type Criteria Considerations (7) Mostly SEG provide more than 10 DOF to reach hand motor function Stroke survivors and Rehabilitation hand disabilities (8) A complete hand characterization must be included Function Grasp, grab, pinch, lift, to tackle more DOF Assistance hold, and release tasks (9) All SEG provide, at least, flexion-extension motion. Cable-driven, pneumatic, Actuation Furthermore, adduction-abduction and opponent or hydraulic motions are desirable Fabric, synthetic leather, Materials neoprene (10) Elastomers have become the main material choice due to their flexibility, lightness, and adaptability Motion guidance Passive or active Mobility Manufacture 3D printing+other procedures (11) SEG systems should have a total mass of less than Operation & control 45 minutes/day, OLC or CLC 200 g to enhance their efficiency 7N/finger or 10 to 15 N Assessment (12) SEG should provide, at least, 5 N per finger to exe- for 1 kg objects cute most of ADL Open or closed palm Modularity (13) Regarding pneumatic actuation, SEG should work configuration between 100 and 300 kPa 500 g or less than 3 kg for Portability the whole system (14) CLC controllers are preferable to OLC in order to Usability Right size, fasteners ensure patient safety and system precision. Particu- Customization or Velcro® straps larly, PD controllers have been mostly implemented Mode of intervention Bilateral or unilateral (15) Fingertip forces, ROM, and EMG are the most used Costs Assembly less than $30 USD tools to evaluate SEG effectiveness Table 3 provides detailed information related to the 15 radial deviations, and pronation-supination; (10) material; aspects illustrated in Figure 2 for 91 devices that have been (11) weight; (12) force; (13) pressure; (14) control: CLC or analysed in order to identify, classify, and discuss the 13 OLC; and (15) assessment. aforementioned criteria and to establish the previous 15 Figure 2 provides information related to the number of design guidelines for SEG systems. For example, the third soft exogloves that have been developed in the last decade, system has eight DOF, focuses on grasping assistance tasks, being characterized by particular aspects. For example, the has a closed palm configuration, is passively driven (CLC) most important number of SEG systems that have been by cables, and performs flexion/extension of 3 fingers. developed is focused on the passive assistance mode, CLC From the previous information reported in this paper, predominate over open-loop strategies, elastomers are pre- five core SEG developers have been identified and have ferred to other types of material, hydraulic actuation is not marked trends in the design of soft exoglove systems. Hong significant compared to the number of SEG devices using Kai Yap is the author with the highest number of SEG contri- tendon-driven or pneumatic actuation, and SEG have been butions (see Table 3, items 25-31). developed, mainly, to cope with stroke and hand disabilities The number of SEG developments, from the last ten as well as with rehabilitation and assistance problems. years, is plotted in Figure 3. According to literature, 2017 Based on what has been presented so far, the following was the most productive year with 21 of the 91 contributions SEG design guidelines are highlighted in order to be consid- reported in this paper. ered when developing new SEG systems. 4. Discussion (1) Rehabilitation and assistance tasks should be included in a single soft exoglove In order to provide technical solutions for hand rehabilita- tion and assistance, multiple endeavours have been done (2) SEG are primarily designed for stroke survivors and during the last three years about SEG developments [128]. people with hand disabilities This review has identified areas of opportunity for the (3) Grasping is the main assistance task that has been improvement of soft exogloves that are used in aided rehabil- addressed by SEG systems itation protocols and assistance tasks. Four main circum- stances have motivated researchers to satisfy popular (4) SEG have been diversified for both OPD and CPD demand and increase SEG development since they represent depending on the actuation an alternative and affordable approach to overcome hand (5) Tendon-driven and pneumatic are preferable types disabilities. These circumstances are related to the increase of actuators in the number of people with hand motor deficits, to poor Applied Bionics and Biomechanics 9 Table 3: Soft exoglove design criterion classification. Palm Assistance Weight Force Pressure SEG # Function Application Task Actuation (number) DOF/finger Fingers Motion Material Control Assessment design mode (g) (N) (kPa) (Ref.) 1 Stroke survivors, Index, middle, Fingertip and AT Grasping CPD Tendon-driven (1) PAM 8 Flexion-extension Synthetic latex 80 18 — CLC [14] SCI thumb blocked force 2 Index, middle, AT-RT Hand disability Grasping CPD Tendon-driven (3) PAM 8 Flexion-extension Synthetic latex —— — CLC (PI, PD) EMG [17] thumb 3 Stroke survivors, Index, middle, AT Grasping CPD Tendon-driven (3) PAM 8 Flexion-extension Synthetic latex 194 20 — CLC EMG [19] SCI thumb 4 Index, middle, Silicone AT Hand disability Grasping OPD Tendon-driven (2) PAM 8 Flexion-extension — 20 —— EMG [21] thumb KE-1300 5 Grasping, Index, middle, Flexion-extension, Silicone AT Hand disability OPD Pneumatic (4) PAM 11 350 22.5 300 CLC ROM, grip strength [23] pinching thumb adduction-abduction KE-1300 T 6 Grasping, Index, middle, Flexion-extension, Silicone AT SCI OPD Tendon-driven (3) PAM 9 104 10.3 — CLC Grip strength [22] pinching thumb opponent KE-1300 T 7 Grasping, Pneumatic All except RT Hand disability OPD PAM 14 Flexion-extension, curl Neoprene 160 —— CLC FEM analysis [8] pinching elastomer (4) thumb 8 Grasping, Hydraulic Flexion-extension, AT-RT Hand disability OPD PAM 15 All Neoprene —— 400 CLC ROM, EMG [51] pinching elastomer (2) opponent 9 Hydraulic Flexion-extension, AT-RT SCI Grasping OPD PAM 15 All Textile — 10-15 — CLC ROM, BBT [11] elastomer (2) opponent 10 Grasping, Hydraulic Flexion-extension, AT Hand disability OPD PAM 15 All Elastomer — 14 413 CLC ROM, EMG [50] holding elastomers opponent Flexion-extension, RT Stroke survivors Gripping OPD Pneumatic PAM 16 All opponent, —— — — — ROM: MI, BBT, FIM [36] adduction-abduction Flexion-extension, 12 Hemiplegic Gripping, AT-RT OPD Pneumatic PAM 16 All opponent, —— — — — ROM: MI, NHPT test [105] patients pinching adduction-abduction Flexion-extension, 13 Gripping, AT-RT Stroke, SCI OPD Pneumatic AAM-PAM 16 All opponent, —— — — — ROM: NHPT, FIM test [45] pinching adduction-abduction Flexion-extension, 14 Gripping, ROM: MI, RT Stroke, SCI OPD Pneumatic AAM-PAM 15 All opponent, —— — — — [113] pinching NHPT test adduction-abduction Flexion-extension, RT Stroke survivors Gripping OPD Electrical (5) PAM 15 All, wrist opponent, —— — — — Ashworth index [38] adduction-abduction AT — Grasping CPD Hydraulic PAM 14 All Flexion-extension — 2620 12 550 CLC Pressure regulating [114] Flexion-extension, 17 Stroke survivors, Grasping, All, wrist, ROM: FMA, OPD Tendon-driven AAM-PAM 14 opponent, —— — — — RT [39] SCI manipulation forearm JTT, PPT pronation-supination Flexion-extension, 18 Grasping, All, wrist, RT Stroke survivors OPD Tendon-driven AAM 14 opponent, —— — — — — [115] manipulation forearm pronation-supination Flexion-extension, 19 Stroke survivors, Grasping, All, wrist, radial-ulnar RT OPD Tendon-driven AAM-PAM 14 Elastomer 132 —— CLC ROM [47] SCI manipulation forearm deviations, pronation- supination 20 Stroke survivors, AT-RT Grasping, pinching CPD Tendon-driven (5) AAM-PAM 12 All Flexion-extension Lycra, fabrics — 15 — CLC Grip force [73] SCI 21 Middle, ring, AT Hand disability Grasping, pinching CPD Tendon-driven (3) PAM 8 Flexion-extension Synthetic leather 700 20 — CLC (PID) Grasping power test [84] thumb 22 Flexion-extension, AT-RT Hand disability Grasping OPD Tendon-driven (5) PAM 15 All Silicone rubber — 17.25 165 CLC ROM: tactile pressure [85] adduction-abduction 10 Applied Bionics and Biomechanics Table 3: Continued. Palm Assistance Weight Force Pressure SEG # Function Application Task Actuation (number) DOF/finger Fingers Motion Material Control Assessment design mode (g) (N) (kPa) (Ref.) 23 Flexion-extension, RT Hand disability Grasping OPD Tendon-driven (5) PAM 15 All Elastomer —— — — — [116] adduction-abduction 24 Material AT Grasping CPD Tendon-driven PAM — All — Fabrics —— — CLC Lift forces [24] handling AT-RT Hand disability Grasping, pinching CPD Pneumatic (5) AAM 14 All Flexion-extension Fabrics 200 — 160 CLC ROM [18] 26 All except RT Stroke survivors Grasping CPD Pneumatic (4) PAM 12 Flexion-extension Fabrics 200 9.25 200 CLC fMRI, optical ROM [41] thumb 27 Grasp Flexion-extension, AT Grasping, releasing CPD Pneumatic (5) PAM 15 All Lycra, fabrics 170 13.6 153 CLC Optical ROM, EMG [43] pathologies opponent AT-RT Stroke survivors Grasping, releasing OPD Pneumatic (5) AAM 14 All Flexion-extension Neoprene 150 — 100 CLC ROM, torque [1] 29 All except AT-RT Stroke survivors Grasping, pinching OPD Pneumatic (4) PAM 12 Flexion-extension Lycra, fabrics 180 10.2 120 CLC fMRI, optical ROM [106] thumb 30 Grasping, lifting, AT Stroke survivors OPD Pneumatic (5) AAM-PAM 14 All Flexion-extension Lycra, fabrics 180 12-36 120 CLC ROM, EMG [44] releasing 31 Grasping, AT-RT Hand disability CPD Pneumatic (5) AAM 14 All Flexion-extension Fabrics 99 13.6 275-375 CLC Gripping force [75] manipulation 32 Index, middle, AT Hand disability Grasping CPD Tendon-driven (1) PAM 8 Flexion-extension Synthetic latex 500 10 — CLC ROM, gripping force [53] thumb 33 Index, middle, AT Muscle weakness Grasping CPD Tendon-driven (2) PAM 8 Flexion-extension Neoprene 1200 10 — CLC (PD) ROM, gripping force [59] thumb, elbow 34 Index, middle, AT Hand disability Grasping CPD Tendon-driven (3) PAM 8 Flexion-extension Neoprene 500 —— CLC (PD) ROM, gripping force [52] thumb AT-RT Heavy tasks Grasping holding CPD Tendon-driven (3) PAM 14 All Flexion-extension Fabrics 770 —— CLC Grasping force [25] 36 Grasping, Pneumatic, All except AT-RT Hand disability OPD PAM 12 Flexion-extension Nylon 150 2.5 230 CLC Bending forces, EMG [30] manipulation hybrid (4) thumb 37 Grasping, Pneumatic, RT Hand disability CPD PAM 14 All Flexion-extension —— — 165.4 CLC ROM, joint angles [48] manipulation hybrid (5) 38 Hand Pneumatic RT Grasping OPD PAM 14 All Flexion-extension Textile —— 526 CLC ROM, fatigue test [42] pathologies elastomers (5) 39 Flexion-extension, AT Hand disability Grasping, pinching CPD Pneumatic (5) PAM 15 All Synthetic leather 135 9 200 CLC ROM, EMG [16] opponent Grasping, RT Stroke survivors releasing, CPD Pneumatic (5) PAM 14 All Flexion-extension Lycra — 15 60 CLC (PID) ROM: FMA, BBT [15] pinching 41 Index, middle, RT Hand disability Grasping CPD Tendon-driven (3) PAM 8 Flexion-extension Synthetic leather —— — CLC ROM, joint angles [31] thumb 42 Index, middle, AT-RT Hand paralysis Grasping CPD Tendon-driven (1) PAM 8 Flexion-extension Polyester fiber 50 35 — CLC ROM, EMG [35] thumb 43 Index, middle, AT-RT Hand disability Gripping CPD Tendon-driven (3) PAM 8 Flexion-extension Fabrics —— — CLC ROM, gripping force [117] thumb 44 Index, middle, AT Older adults Gripping CPD Tendon-driven (3) PAM 8 Flexion-extension Fabrics 85 —— CLC Pinch strength, JTHFT [118] thumb 45 Pinching force, AT-RT Hand disability Gripping CPD Tendon-driven (5) PAM 14 All Flexion-extension —— — — CLC [46] JTHFT test RT Hand disability Grasping CPD Tendon-driven (5) PAM 14 All Flexion-extension Fabrics —— — CLC ROM, joint angles [54] Flexion-extension, RT Stroke survivors Grasping OPD Pneumatic (5) PAM 16 All opponent, Nylon — 20 — CLC (PID) ROM, joint angles [71] adduction-abduction Applied Bionics and Biomechanics 11 Table 3: Continued. Palm Assistance Weight Force Pressure SEG # Function Application Task Actuation (number) DOF/finger Fingers Motion Material Control Assessment design mode (g) (N) (kPa) (Ref.) 48 Spring RT Stroke survivors Grasping CPD PAM 14 All Flexion-extension Synthetic leather 200 22.59 —— Bending force [34] mechanism (5) RT Stroke survivors Grasping OPD Pneumatic (5) PAM 14 All Flexion-extension —— — — CLC Electroencephalography [110] EMG, ROM, AT-RT Stroke survivors Gripping OPD Pneumatic (5) PAM 14 All Flexion-extension Elastomer — 41.8 200 CLC [89] gripping force 51 Grasping, AT Hand disability CPD Cable-driven (4) PAM 11 All except little Flexion-extension — 250 16 — CLC ROM, pinching force [107] manipulation 52 All except RT Hand disability Grasping CPD Pneumatic (4) PAM 12 Flexion-extension Nylon — 3 200 — FEM & ROM [67] thumb 53 Grasping, Pneumatic artificial AT Hand disability OPD PAM 14 All Flexion-extension Fabrics 161 10 200 — FEM, fingertip force [119] manipulation muscles (5) 54 Grasping, TPU, Pressure regulation AT-RT Older adults OPD Tendon-driven (5) PAM 14 All Flexion-extension 50 40 60 CLC [120] manipulation NINJAFLEX™ & fingertip force 55 22 pinch, 48 Pinching and AT-RT Hand disability Grasping, pinching OPD Tendon-driven (5) PAM 14 All Flexion-extension TPU 330 — CLC [74] grasp grasping forces 56 Shape memory ROM, fingerti p-tendon AT-RT Hand disability Grasping CPD PAM 14 All Flexion-extension Fabrics — 40 — CLC [55] alloys (5) force 57 Grasping, All except AT Stroke survivors OCP Tendon-driven (5) PAM 12 Flexion-extension Polymer 340 —— CLC Gripping force [4] manipulation thumb 58 CP, stroke Thumb, index, Grasping and AT Grasping CPD Tendon-driven (3) PAM 8 Flexion-extension Synthetic leather 55 48 — CLC [112] survivors and middle fingertip forces Flexion-extension, 59 Shape memory radial abduction, AT Supportive aid Grasping CPD PAM 12 All except little Synthetic leather 85.03 11 — CLC Grasping force, ROM [121] alloys (5) palmar abduction, opposition RT Stroke survivors Grasping OPD Pneumatic (1) PAM 3 Index Flexion-extension Ecoflex™ 00-30 — 1.17 30 CLC Bending angle [122] 61 Older adults, Pneumatic artificial Grasping and AT Grasping, pinching CPD PAM 14 All Flexion-extension Rubber — 5.7-14, 20-25 500 CLC [13] hand disability muscles (5) pinching forces 62 Thumb, index, Writing tasks, AT SCI Grasping CPD Tendon-driven (3) PAM 8 Flexion-extension —— 7.39 — CLC [109] and middle grasping force 63 Pneumatic, AT-RT Stroke survivors Grasping CPD PAM 14 All Flexion-extension Nylon — 5 300 CLC Ashworth test [7] tendon-driven (5) 64 All except RT Stroke survivors Grasping CPD Tendon-driven (4) PAM 12 Flexion-extension —— — — CLC FMA test [123] thumb 65 Pneumatic RT Hand disability Grasping OPD PAM 14 All Flexion-extension Nylon 280 11.27 250 CLC Grasping test FEM [82] FREAs (5) 66 Pneumatic RT Hand disability Grasping OPD PAM 1 Thumb Opposition Elastomer 586 — 150 CLC Kapandji test [108] FREAs (1) 67 Pneumatic AT Stroke survivors Grasping OPD PAM 14 All Flexion-extension Silicone rubber 207 — 200 CLC Bending force FEM [124] FREAs (5) 68 Grasping, AT SCI OPD Pneumatic (5) PAM 14 All Flexion-extension Fabric 77 15 172 CLC Lifting force [125] manipulation AT Stroke survivors Grasping CPD Electrical PAM 15 All, wrist Flexion-extension Neoprene fabric —— — CLC BBT test [99] RT Stroke survivors Grasping CPD Tendon-driven (1) PAM 3 Index, wrist Flexion-extension Lycra —— — CLC ROM [98] 71 Kinesthetics, Index and Virtual reality RT Pressing OPD Pneumatic (2) PAM 6 Flexion Silicone rubber — 16.66 210 CLC [61] haptic feedbacks middle haptic feedback 72 Grasping, Flexion-extension, AT Hand disabilities CPD Pneumatic (4) PAM 15 All — 65 —— CLC Grabbing force [126] holding opposition 12 Applied Bionics and Biomechanics Table 3: Continued. Palm Assistance Weight Force Pressure SEG # Function Application Task Actuation (number) DOF/finger Fingers Motion Material Control Assessment design mode (g) (N) (kPa) (Ref.) AT Hand disabilities Grasping CPD Pneumatic (4) PAM 11 All except little Flexion-extension — 160 25 500 CLC Bending angle [57] AT Hand disabilities Grasping CPD Pneumatic (5) PAM 14 All Flexion-extension Elastomer 180 3 300 CLC Grasping force [127] 75 CLC RT Stroke survivors Grasping CPD Pneumatic (5) PAM 14 All Flexion-extension Latex —— — Bending angle [68] proportional 76 Thumb, index, Flexion-extension, RT Stroke survivors Grasping OPD Tendon-driven (5) PAM 9 Silicone KE-1300 T 120 12 — CLC Bending angle [72] and middle opposition/reposition 77 All except AT Stroke survivors Grasping CPD Tendon-driven (5) PAM 12 Flexion-extension Lycra — 16-17 — CLC Fingertip force ROM [49] thumb RT Stroke survivors Grasping CPD Tendon-driven (5) AAM-PAM 14 All Flexion-extension Elastomer >1000 —— CLC (PD) Fingertip force ROM [100] Flexion-extension, 79 Pneumatic Bending angle and AT-RT Hand disabilities Grasping CPD PAM 16 All opponent, adduction- Polyester 76 0.8 150 CLC [69] FREA (5) force output abduction 80 RTV-4234T4, RT Stroke survivors Grasping OPD Pneumatic (5) PAM 14 All Flexion-extension —— 50 CLC (PD) Bending angle [88] silicon RT Stroke survivors Grasping CPD Tendon-driven (5) PAM 14 All Flexion-extension Fabrics —— — — Virtual reality, FMA [40] AT Hand disabilities Grasping Semiopen Motor-tendon (5) PAM 14 All Flexion-extension Cotton fabric 600 —— On-off control Grasping force output [10] 83 Flexion-extension, AT Hand disabilities Grasping CPD Tendon-driven (5) PAM 12 All except little — 220 83 — CLC Grasping force ROM [90] opponent 84 Tendon-driven AT Heavy tasks Manipulation tasks CPD PAM 14 All Flexion-extension Rubber — 70 — CLC (PID) Force output [26] SMA (5) AT Hand disabilities Grasping, releasing CPD Pneumatic (5) PAM 14 All Flexion-extension Fabric 160 88.29 180 — EMG signals [104] 86 ROM, force output, RT Hand disabilities Grasping CPD Steel spring AAM-PAM 14 All Flexion-extension — 401 30.87 — CLC [91] EEG signals 87 EMG signals, grasping RT Hand disabilities Grasping CPD Pneumatic (5) PAM 14 All Flexion-extension — 150 40 300 OLC [58] forces 88 EMG signals, grasping AT Stroke survivors Grasping CPD Tendon-driven (5) AAM-PAM 10 All Flexion-extension Nylon 258 —— OLC [96] and lifting forces 89 Tendon-driven Teleoperation, RT Stroke survivors Grasping CPD PAM 14 All Flexion-extension —— — — — [103] SMA (5) time output 90 RTV-4234T4, Finger trajectories RT Stroke survivors Grasping CPD Pneumatic (5) PAM 14 All Flexion-extension —— 105 CLC (PD) [101] silicon and angle 91 Polyester and Griping force and AT Hand disabilities Grasping CPD Tendon-driven (5) PAM 14 All Flexion-extension —— — CLC [20] neoprene FMG signals Applied Bionics and Biomechanics 13 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 Year Figure 3: SEG developments in the last decade. adults. Thus, adjustable devices are recommended to have therapist availability, to the fact that clinical facilities are struggling to provide rehabilitation training, and to the the possibility to initiate an early SEG-based rehabilitation expensive costs of these health services. program since this is a common advice given by therapists, There are still significant challenges to face in soft exoglove no matter the dimensions of the patient’s hand. So far, SEG design. For instance, power supply approaches are still limited systems are able to accomplish full open-close fist, grasping, and tendon-driven actuation necessitates motors without lifting, and object release. Therefore, the systems reported heating problems, whereas hysteresis issues should be solved in literature encompass from 8 to 14 DOF. Moreover, SEG in pneumatic systems to increase actuation cycles and durabil- characterization could be developed to obtain more DOF in ity along with lightweight and portable power supplies. order to expand the workspace if needed. Regarding rehabilitation approaches, SEG systems must When soft exogloves are used, patient safety is a priority. be endowed to exert intensive and repetitive routines without Thus, human-machine interfaces with emergency buttons muscle fatigue and with minimal therapist assistance to excel and haptic feedback must be considered for harmless interac- above other rehabilitation options. SEG are a supportive aid tions [35, 128] as stated in Section 2 of this paper, and several safety strategies must be incorporated in every SEG system. that contributes to accelerated hand recovery by therapy pro- tocols. Nevertheless, to achieve a desired rehabilitation task, Moreover, SEG systems should not obstruct natural hand an active contribution from the patient is required to regain mobility and do not affect active ROM. Additionally, new strength, mobility, and ROM. Since the progress of each developments are expected to provide patients and thera- patient is variable, an AAM with time-triggered control could pists with useful information in order to evaluate patient progress. Furthermore, the capability to automatically adjust be implemented to regulate the input force of patients during rehabilitation processes, depending on their physical condi- the operation parameters as a function of the patient recov- tion. SEG systems must encourage patient participation but ery level is desirable. do not execute all the rehabilitation work. SEG self-manufacturing designs must ensure functional Several works have demonstrated that soft exogloves operation for home rehabilitation to provide low-cost sys- tems. These considerations could allow to improve SEG fea- have the potential to offer safe human-robot rehabilitation or assistance. However, new trends show that these two tasks tures as hours of operation, power consumption, cleaning, should be integrated into a unified system as it is reported by and maintenance. Since Bluetooth communications have [46, 92]. To accomplish integral rehabilitation, SEG designers been considered between SEG systems and control interfaces must consider that modular devices are expected to help [74], this or other communication systems must be part of new SEG devices when dealing with CLC strategies and for therapists and patients depending on the impairment or on the rehabilitation protocol. This will be satisfied by connect- rehabilitation or assistance data analysis. ing a soft exoglove device to a soft exosuit with a reliable and From this review, it can be pointed out that in recent robust platform (see, for instance, [28]). years, the development of SEG has grown significantly in SEG shortcomings were identified concerning different rehabilitation clinics and research groups. However, there is no comparison between research prototypes and those that hand sizes since most available systems are oriented towards Number of soft exogloves 14 Applied Bionics and Biomechanics criteria have been identified, classified, and established into have been already commercialized because the level of their technological maturity is different for each of them. Com- 2 function, 6 operation, and 5 usability criteria. mercialized SEG systems must have evolved from research This paper also provides 15 guidelines for SEG design, a prototypes. The main difference between these two types of detailed description of 91 SEG that have been analysed based devices is the one related to their technological maturity. on the aforementioned criteria, and a discussion that con- For instance, research prototypes can reach, in favorable siders different aspects in order to enhance future SEG cases, a technology readiness level (TRL) of 4 or 5 while com- developments. mercialized products have the highest TRL of 9 in China From this review, it is highlighted that patient safety [129, 130]. The evolution of a research prototype going from should be a priority characteristic during SEG operation, a 5 TRL to a certified product with 8 TRL and to a commer- and then, it should be guaranteed in every new SEG devel- cial product with a 9 TRL can take several years and require opment. This goal can be achieved by working closely with significant quantities of money. Moreover, medical devices a therapist, as recommended in [28], as well as incorporat- having official approvals or certifications as that of the Food ing safety in mechanical and electronical parts and in the and Drug Administration (FDA) or the Conformité Europé- programming of the SEG device. Moreover, safety stan- enne (CE) can be commercialized since they satisfy specific dards have been referenced to be considered in every requirements and standards while research prototypes focus, SEG development. mainly, on satisfying functional aspects. Then, it can be It has been remarked that several efforts have been made stated that commercialized medical devices are reliable due in terms of SEG designs. However, there is still room to to the fact that they have completed the product design cycle improve these devices. Then, this paper provides suggestions reaching the product life-cycle management, while research on patient safety, functional and continuous operation, prototypes have not begun the product development cycle friendly interaction, feedback information, and materials. or their industrial manufacture yet. Other areas to be explored include hybrid SEG systems New-generation products should seek for an affordable where new assembly techniques ensure force transmission trade-off between cost and benefit and include the possibility or the use of electroencephalography signals to monitor to perform assistance or rehabilitation therapy at home or in brain activity when SEG rehabilitation is performed. SEG specialized clinics to ensure that rehabilitation protocols, systems should be able to combine passive and active assis- defined by therapists, are efficiently executed. tance modes along with bilateral training to enhance recov- SEG designs should provide acceptable appearance, com- ery processes and to encourage patients. The mentioned fort, and functionality to patients. Hence, it is highly recom- SEG design criteria provide perfectible guidelines to improve mended that SEG systems consider accessible technologies their performance and represent a basis to develop SEG that could, additionally, create dynamic environments where robust designs. patients can have pleasant therapy sessions. SEG require materials with appearance and elastic modulus similar to Abbreviations human tissues. Thus, smart polymers represent the primary current choice due to their biomimetic qualities to develop AAM: Active assistance mode lightweight devices with modular OPD [128]. Besides, elasto- ADL: Activities of daily living mers have been shown to be compliant wearable components ARM: Active resistive mode with the ability to vary their form and increase the ROM ASI: Ashworth spasticity index based on the shape of the human hand. AT: Assistance tasks Modularity plays a significant role when dealing with BBT: Box and block test maintenance aspects of SEG systems as well as with costs CLC: Closed-loop control and should be considered in new SEG developments. Besides, CE: Conformité Européenne modularity can play a significant role when dealing with CMC: Carpometacarpal rehabilitation of different fingers or DOF. Regarding porta- CP: Cerebral palsy bility in new SEG developments, minimizing the dependence CPD: Closed palm design of energy sources becomes a challenge that must be DOF: Degrees of freedom addressed by researchers and engineers. EMG: Electromyography It has become clear that a SEG device that allows adapta- FDA: Food and Drug Administration tion (customization) to a larger number of patients without FEM: Finite element method the need for component replacements will be preferable to FIM: Functional independence measure another system that only works for a certain size of hands. FMA: Fugl-Meyer assessment FMG: Force myography FREA: Fiber reinforced elastomer actuators 5. Conclusions GUI: Graphical user interface IP: Interphalangeal JTT: Jebsen-Taylor hand test Scientific and technical communications concerning wear- able SEG for hand rehabilitation and assistance tasks applied MAS: Modified Ashworth scale to stroke survivors or people with hand disabilities have been MCP: Metacarpophalangeal extensively reviewed and reported in this paper. SEG design MIT: Motricity index test Applied Bionics and Biomechanics 15 [9] Z. Yue, X. Zhang, and J. Wang, “Hand Rehabilitation Robot- NHPT: Nine-hole peg test ics on Poststroke Motor Recovery,” Behavioural Neurology, OLC: Open-loop control vol. 2017, Article ID 3908135, 20 pages, 2017. OPD: Open palm design [10] R. Ismail, M. Ariyanto, T. Hidayat, and J. D. Setiawan, PAM: Passive assistance mode “Design of fabric-based soft robotic glove for hand function PD: Proportional-derivative assistance,” in 2019 6th International Conference on Informa- PID: Proportional-integral-derivative tion Technology, Computer and Electrical Engineering (ICI- PT: Purdue pegboard test TACEE), pp. 1–5, Semarang, Indonesia, Indonesia, 2019. ROM: Range of motion [11] P. Polygerinos, K. C. Galloway, E. Savage, M. Herman, RT: Robot rehabilitation K. O’Donnell, and C. J. Walsh, “Soft robotic glove for hand SCI: Spinal cord injury rehabilitation and task specific training,” in 2015 IEEE Inter- SEG: Soft exogloves national Conference on Robotics and Automation (ICRA), SMA: Shape memory alloys pp. 2913–2919, Seattle, WA, USA, 2015. RTV: Room-temperature-vulcanizing [12] D. Rus and M. T. Tolley, “Design, fabrication and control of TPU: Thermoplastic polyurethane soft robots,” Nature, vol. 521, no. 7553, pp. 467–475, 2015. TRL: Technology readiness level. [13] T. Noritsugu, M. Takaiwa, and D. Sasaki, “Power assist wear driven with pneumatic rubber artificial muscles,” in 2008 Conflicts of Interest 15th International Conference on Mechatronics and Machine Vision in Practice, pp. 539–544, Auckland, New Zealand, The authors declare that there is no conflict of interest regarding the publication of this paper. [14] H. In, K. J. Cho, K. Kim, and B. Lee, “Jointless structure and under-actuation mechanism for compact hand exoskeleton,” Acknowledgments in 2011 IEEE International Conference on Rehabilitation Robotics, pp. 1–6, Zurich, Switzerland, 2011. The authors thank CONACYT and Universidad Autón- [15] L. Connelly, Y. Jia, M. L. Toro, M. E. Stoykov, R. V. Kenyon, oma del Estado de México for all their support. Funding and D. G. Kamper, “A Pneumatic Glove and Immersive Vir- supporting this research was provided by CONACYT tual Reality Environment for Hand Rehabilitative Training (406476) and Universidad Autónoma del Estado de México After Stroke,” IEEE Transactions on Neural Systems and (5015-2020CIA). Rehabilitation Engineering, vol. 18, no. 5, pp. 551–559, 2010. [16] Y. Kadowaki, T. Noritsugu, M. Takaiwa, D. Sasaki, and References M. Kato, “Development of Soft Power-Assist Glove and Con- trol Based on Human Intent,” Journal of Robotics and Mecha- [1] H. K. Yap, J. H. Lim, J. C. H. Goh, and C.-H. Yeow, “Design of tronics, vol. 23, no. 2, pp. 281–291, 2011. a Soft Robotic Glove for Hand Rehabilitation of Stroke [17] U. Jeong, H.-K. In, and K.-J. Cho, “Implementation of various Patients With Clenched Fist Deformity Using Inflatable Plas- control algorithms for hand rehabilitation exercise using tic Actuators,” Journal of Medical Devices, vol. 10, no. 4, 2016. wearable robotic hand,” Intelligent Service Robotics, vol. 6, [2] V. L. Feigin, B. Norrving, and G. A. Mensah, “Global burden no. 4, pp. 181–189, 2013. of stroke,” Circulation Research, vol. 120, no. 3, pp. 439–448, 2017. [18] H. K. Yap, J. H. Lim, F. Nasrallah, J. C. H. Goh, and R. C. H. Yeow, “A soft exoskeleton for hand assistive and reha- [3] A. G. Thrift, T. Thayabaranathan, G. Howard et al., “Global bilitation application using pneumatic actuators with vari- stroke statistics,” International Journal of Stroke, vol. 12, able stiffness,” in 2015 IEEE International Conference on no. 1, pp. 13–32, 2017. Robotics and Automation (ICRA), pp. 4967–4972, Seattle, [4] B. Cherian, C. Dominic, G. Vysakh, and K. R. Vishakh, “Exo- WA, USA, 2015. Glove: A Soft Wearable Robotic Hand for Stroke Survivors,” [19] H. In, B. B. Kang, M. Sin, and K. J. Cho, “Exo-Glove: A Wear- International Research Journal of Engineering and Technol- able Robot for the Hand with a Soft Tendon Routing System,” ogy, vol. 5, no. 5, pp. 3358–3364, 2018. IEEE Robotics and Automation Magazine, vol. 22, no. 1, [5] B. Radder, G. B. Prange-Lasonder, A. I. R. Kottink et al., pp. 97–105, 2015. “Home rehabilitation supported by a wearable soft-robotic [20] J. Fajardo, A. R. Neto, W. Silva, M. Gomes, E. Fujiwara, and device for improving hand function in older adults: A pilot randomized controlled trial,” PLoS One, vol. 14, no. 8, article E. Rohmer, “A wearable robotic glove based on optical FMG driven controller,” in 2019 IEEE 4th International Con- e0220544, 2019. ference on Advanced Robotics and Mechatronics (ICARM), [6] B. Radder, “General Introduction,” The Wearable Hand pp. 81–86, Toyonaka, Japan, Japan, 2019. Robot, pp. 10–21, 2018. [21] B. B. Kang, H. Lee, H. In, U. Jeong, J. Chung, and K. J. Cho, [7] A. Stilli, A. Cremoni, M. Bianchi et al., “AirExGlove — A “Development of a polymer-based tendon-driven wearable novel pneumatic exoskeleton glove for adaptive hand rehabil- robotic hand,” in 2016 IEEE International Conference on itation in post-stroke patients,” in IEEE International Confer- Robotics and Automation (ICRA), pp. 3750–3755, Stockholm, ence on Soft Robotics (RoboSoft), vol. 2018, pp. 579–584, Sweden, 2016. Livorno, Italy, 2018. [8] P. Polygerinos, S. Lyne, Z. Wang et al., “Towards a soft pneu- [22] B. B. Kang, H. Choi, H. Lee, and K. J. Cho, “Exo-Glove Poly matic glove for hand rehabilitation,” in 2013 IEEE/RSJ Inter- II: A Polymer-Based Soft Wearable Robot for the Hand with national Conference on Intelligent Robots and Systems, a Tendon-Driven Actuation System,” Soft Robotics, vol. 6, pp. 1512–1517, Tokyo, Japan, 2013. no. 2, pp. 214–227, 2019. 16 Applied Bionics and Biomechanics [38] P. Bernocchi, C. Mulè, F. Vanoglio, G. Taveggia, A. Luisa, and [23] S. S. Yun, B. B. Kang, and K. J. Cho, “Exo-Glove PM: An Eas- ily Customizable Modularized Pneumatic Assistive Glove,” S. Scalvini, “Home-based hand rehabilitation with a robotic IEEE Robotics and Automation Letters, vol. 2, no. 3, glove in hemiplegic patients after stroke: a pilot feasibility pp. 1725–1732, 2017. study,” Topics in Stroke Rehabilitation, vol. 25, pp. 114–119, [24] H. Kazerooni, D. Fairbanks, A. Chen, and G. Shin, “The magic glove,” in 2004 in IEEE International Conference on [39] J.-H. Shin, M.-Y. Kim, J.-Y. Lee et al., “Effects of virtual reality-based rehabilitation on distal upper extremity func- Robotics and Automation, 2004. Proceedings. ICRA '04. 2004, vol. 1, pp. 757–763, New Orleans, LA, USA, USA, 2004. tion and health-related quality of life: a single-blinded, ran- domized controlled trial,” Journal of NeuroEngineering and [25] M. Diftler, C. A. Ihrke, L. B. Bridgwater et al., “RoboGlove - A Rehabilitation, vol. 13, no. 1, 2016. Robonaut Derived Multipurpose Assistive Device,” in Inter- [40] B. Radder, G. B. Prange-Lasonder, A. I. R. Kottink, national Conference on Robotics and Automation, Hong A. Melendez-Calderon, J. H. Buurke, and J. S. Rietman, “Fea- Kong, China, 2014. sibility of a wearable soft-robotic glove to support impaired [26] Á. Villoslada, C. Rivera, N. Escudero, F. Martín, D. Blanco, hand function in stroke patients,” Journal of Rehabilitation and L. Moreno, “Hand Exo-Muscular System for Assisting Medicine, vol. 50, no. 7, pp. 598–606, 2018. Astronauts During Extravehicular Activities,” Soft Robotics, [41] H. K. Yap, J. H. Lim, F. Nasrallah, F. Z. Low, J. C. H. Goh, and vol. 6, no. 1, pp. 21–37, 2019. R. C. H. Yeow, “MRC-glove: A fMRI compatible soft robotic [27] P. Ohta, L. Valle, J. King et al., “Design of a Lightweight glove for hand rehabilitation application,” in 2015 IEEE Inter- Soft Robotic Arm Using Pneumatic Artificial Muscles and national Conference on Rehabilitation Robotics (ICORR), Inflatable Sleeves,” Soft Robotics, vol. 5, no. 2, pp. 204– pp. 735–740, Singapore, Singapore, 2015. 215, 2018. [42] Y. Jiang, D. Chen, J. Que, Z. Liu, Z. Wang, and Y. Xu, “Soft [28] J.-M. Dávila-Vilchis, J. C. Á. V. LAZ-Avilés, and A. H. robotic glove for hand rehabilitation based on a novel fabrica- Vilchis-González, “Design Methodology for Soft Wearable tion method,” in 2017 IEEE International Conference on Devices—The MOSAR Case,” Applied Sciences, vol. 9, no. 22, Robotics and Biomimetics (ROBIO), pp. 817–822, Macau, p. 4727, 2019. China, 2017. [29] M. Troncossi, M. Mozaffari-Foumashi, and V. Parenti-Cas- [43] H. K. Yap, B. W. K. Ang, J. H. Lim, J. C. H. Goh, and C. H. telli, “An original classification of rehabilitation hand exo- Yeow, “A fabric-regulated soft robotic glove with user intent skeletons,” Journal of Robotics and Mechanical Engineering detection using EMG and RFID for hand assistive application,” Research, vol. 1, no. 4, pp. 17–29, 2016. in 2016 IEEE International Conference on Robotics and Auto- [30] Y. Chen, S. Le, Q. C. Tan, O. Lau, F. Wan, and C. Song, “A mation (ICRA),pp.3537–3542, Stockholm, Sweden, 2016. lobster-inspired robotic glove for hand rehabilitation,” in [44] H. K. Yap, J. H. Lim, F. Nasrallah, and C.-H. Yeow, “Design 2017 IEEE International Conference on Robotics and Automa- and Preliminary Feasibility Study of a Soft Robotic Glove tion (ICRA), pp. 4782–4787, Singapore, Singapore, 2017. for Hand Function Assistance in Stroke Survivors,” Frontiers [31] S. Biggar and W. Yao, “Design and Evaluation of a Soft and in Neuroscience, vol. 11, 2017. Wearable Robotic Glove for Hand Rehabilitation,” IEEE [45] F. Vanoglio, A. Luisa, F. Garofali, and C. Mora, “Evaluation Transactions on Neural Systems and Rehabilitation Engineer- of the effectiviness of Glohera (hand rehabilitation glove) ing, vol. 24, no. 10, pp. 1071–1080, 2016. on hemiplegic patients. Pilot study,” in XIII Congress of Ital- [32] J. Brackenridge, L. V. Bradnam, S. Lennon, J. J. Costi, and ian Society of Neurorehabilitation, pp. 18–20, Bari, Italy, D. A. Hobbs, “A Review of Rehabilitation Devices to Promote Upper Limb Function Following Stroke,” Neuroscience and [46] G. B. Prange-Lasonder, B. Radder, A. I. R. Kottink, Biomedical Engineering, vol. 4, no. 1, pp. 25–42, 2016. A. Melendez-Calderon, J. H. Buurke, and J. S. Rietman, [33] G. A. Prieto, A. I. Andrés, A. Venkatakrishnan, W. Q. Malik, “Applying a soft-robotic glove as assistive device and training V. Dietz, and W. Z. Rymer, “Rehabilitation Technologies for tool with games to support hand function after stroke: Pre- Spinal Injury,” in Emerging Therapies in Neurorehabilitation liminary results on feasibility and potential clinical impact,” II, pp. 65–85, Springer, 2016. in 2017 International Conference on Rehabilitation Robotics [34] K. Serbest, S. Ateş, and A. H. A. Stienen, “Design of an exer- (ICORR), pp. 1401–1406, London, UK, 2017. cise glove for hand rehabilitation using spring mechanism,” [47] H. Kim, A. Lee, Y. I. Shin et al., “Effects of digital smart glove in 2016 20th National Biomedical Engineering Meeting system on motor recovery of upper extremity in subacute (BIYOMUT), pp. 1–5, Izmir, Turkey, 2016. stroke patients,” Annals of Physical and Rehabilitation Medi- cine, vol. 61, p. e28, 2018. [35] H. Cao and D. Zhang, “Soft robotic glove with integrated sEMG sensing for disabled people with hand paralysis,” in [48] M. Haghshenas-Jaryani, W. Carrigan, C. Nothnagle, and M. B. 2016 IEEE International Conference on Robotics and Biomi- J. Wijesundara, “Sensorized soft robotic glove for continuous metics (ROBIO), pp. 714–718, Qingdao, China, 2016. passive motion therapy,” in 2016 6th IEEE International Con- ference on Biomedical Robotics and Biomechatronics (BioRob), [36] A. Rodigari, A. Zancan, I. Jedrychowska, and I. Springhetti, pp. 815–820, Singapore, Singapore, 2016. “Effectiveness of the treatment of hemiplegic patient’s hand with robotic rehabilitation glove “gloreha”: preliminary [49] S. Toochinda and W. Wannasuphoprasit, “Design and results,” in The 42nd National Congress Of The Italian Society Development of an Assistive Hand Device for Enhancing Of Physical And Rehabilitative Medicine, pp. 142–145, Compatibility and Comfortability,” in 2018 2nd International Torino, Italy, 2014, SIMFER. Conference on Engineering Innovation (ICEI), pp. 1–6, Bang- kok, Thailand, 2018. [37] P. Maeder-York, T. Clites, E. Boggs et al., “Biologically Inspired Soft Robot for Thumb Rehabilitation1,” Journal of [50] P. Polygerinos, K. C. Galloway, S. Sanan, M. Herman, and Medical Devices, vol. 8, no. 2, 2014. C. J. Walsh, “EMG controlled soft robotic glove for assistance Applied Bionics and Biomechanics 17 during activities of daily living,” in 2015 IEEE International [65] Y. Tsuji, “A safety review of medical device robots in Japan,” Conference on Rehabilitation Robotics (ICORR), pp. 55–60, sia Pacific Journal of Health Law &Ethics, vol. 12, no. 1, pp. 1– Singapore, Singapore, 2015. 27, 2018. [51] P. Polygerinos, Z. Wang, K. C. Galloway, R. J. Wood, and C. J. [66] P. Barattini, V. Federico, G. S. Virk, and T. Haidegger, Human–Robot Interaction, Taylor & F. Boca Raton: CRC Walsh, “Soft robotic glove for combined assistance and at- home rehabilitation,” Robotics and Autonomous Systems, Press, New York NY USA, 2019. vol. 73, pp. 135–143, 2015. [67] Z. Sun, Z. Guo, and W. Tang, “Design of wearable hand reha- bilitation glove with soft hoop-reinforced pneumatic actua- [52] M. Xiloyannis, L. Cappello, B. Khanh Dinh, C. W. Antuvan, tor,” Journal of Central South University, vol. 26, no. 1, and L. Masia, “Design and Preliminary Testing of a Soft Exo- pp. 106–119, 2019. suit for Assisting Elbow Movements and Hand Grasping,” in Converging Clinical and Engineering Research on Neuroreh- [68] H. Li and L. Cheng, “Preliminary study on the design and abilitation II, vol. 15, pp. 557–561, Springer, 2017. control of a pneumatically-actuated hand rehabilitation device,” in 2017 32nd Youth Academic Annual Conference [53] M. Xiloyannis, L. Cappello, D. B. Khanh, S.-C. Yen, and of Chinese Association of Automation (YAC), pp. 860–865, L. Masia, “Modelling and design of a synergy-based actuator Hefei, 2017. for a tendon-driven soft robotic glove,” in 2016 6th IEEE [69] T. Jiralerspong, K. H. L. Heung, R. K. Y. Tong, and Z. Li, “A International Conference on Biomedical Robotics and Biome- chatronics (BioRob), pp. 1213–1219, Singapore, Singapore, novel soft robotic glove for daily life assistance,” in 2018 7th IEEE International Conference on Biomedical Robotics and Biomechatronics (Biorob), pp. 671–676, Enschede, 2018. [54] C. Jiménez, R. Mora, J. Pérez, and K. Quirós, “Nitiglove: [70] L. Wang, S. G. Nurzaman, and F. Iida, “Soft-material robot- Nitinol-Driven Robotic Glove Used to Assist Therapy for ics,” Foundations and Trends in Robotics, vol. 5, no. 3, Hand Mobility Recovery,” in Wearable Robotics: Challenges pp. 191–259, 2017. and Trends, pp. 379–383, Springer, 2017. [71] B. Wang, A. McDaid, M. Biglari-Abhari, and K. C. Aw, [55] A. Hadi, K. Alipour, S. Kazeminasab, and M. Elahinia, “ASR “Design and development of a glove for post-stroke hand glove: A wearable glove for hand assistance and rehabilitation rehabilitation,” in 2017 IEEE International Conference on using shape memory alloys,” Journal of Intelligent Material Advanced Intelligent Mechatronics (AIM), pp. 1047–1051, Systems and Structures, vol. 29, no. 8, pp. 1575–1585, 2017. Munich, 2017. [56] S. Sridar, P. H. Nguyen, M. Zhu, Q. P. Lam, and [72] D. H. Kim and H. S. Park, “Cable actuated dexterous P. Polygerinos, “Development of a soft-inflatable exosuit for (CADEX) glove for effective rehabilitation of the hand for knee rehabilitation,” in 2017 IEEE/RSJ International Confer- patients with neurological diseases,” in 2018 IEEE/RSJ Inter- ence on Intelligent Robots and Systems (IROS), pp. 3722– national Conference on Intelligent Robots and Systems (IROS), 3727, Vancouver, BC, Canada, 2017. Madrid, Spain, 2018. [57] H. Al-Fahaam, S. Davis, S. Nefti-Meziani, and T. Theodoridis, [73] M. A. Delph, S. A. Fischer, P. W. Gauthier, C. H. M. Luna, “Novel soft bending actuator-based power augmentation hand E. A. Clancy, and G. S. Fischer, “A soft robotic exomuscula- exoskeleton controlled by human intention,” Intelligent Service ture glove with integrated sEMG sensing for hand rehabilita- Robotics,vol. 11,no.3, pp.247–268, 2018. tion,” in 2013 IEEE 13th International Conference on [58] H. Al-Fahaam, S. Davis, and S. Nefti-Meziani, “The design Rehabilitation Robotics (ICORR), pp. 1–7, Seattle, WA, 2013. and mathematical modelling of novel extensor bending [74] A. Mohammadi, J. Lavranos, P. Choong, and D. Oetomo, pneumatic artificial muscles (EBPAMs) for soft exoskele- “Flexo-glove: a 3D printed soft exoskeleton robotic glove for tons,” Robotics and Autonomous Systems, vol. 99, pp. 63–74, impaired hand rehabilitation and assistance,” in 2018 40th Annual International Conference of the IEEE Engineering [59] M. Xiloyannis, L. Cappello, K. D. Binh, C. W. Antuvan, and in Medicine and Biology Society (EMBC), pp. 2120–2123, L. Masia, “Preliminary design and control of a soft exosuit Honolulu, HI, 2018. for assisting elbow movements and hand grasping in activi- [75] H. K. Yap, P. M. Khin, T. H. Koh et al., “A fully fabric-based ties of daily living,” Journal of Rehabilitation and Assistive, bidirectional soft robotic glove for assistance and rehabilita- vol. 4, article 205566831668031, 2017. tion of hand impaired patients,” IEEE Robotics and Automa- [60] H. Abidi and M. Cianchetti, “On intrinsic safety of soft tion Letters, vol. 2, no. 3, pp. 1383–1390, 2017. robots,” Frontiers in Robotics and AI, vol. 4, no. 5, 2017. [76] H. K. Yap, J. C. H. Goh, and R. C. H. Yeow, “Design and char- [61] S. Jadhav, V. Kannanda, B. Kang, M. T. Tolley, and J. P. acterization of soft actuator for hand rehabilitation applica- Schulze, “Soft robotic glove for kinesthetic haptic feedback tion,” in 6th European Conference of the International in virtual reality environments,” Electronic Imaging, Federation for Medical and Biological Engineering, Springer, vol. 2017, no. 3, pp. 19–24, 2017. pp. 367–370, Dubrovnik, 2015. [62] A. Avižienis, J. C. Laprie, B. Randell, and C. Landwehr, “Basic [77] H. K. Yap, J. H. Lim, F. Nasrallah, J. C. H. Goh, and C.- concepts and taxonomy of dependable and secure comput- H. Yeow, “Characterisation and evaluation of soft elastomeric ing,” IEEE Transactions on Dependable and Secure Comput- actuators for hand assistive and rehabilitation applications,” ing, vol. 1, no. 1, pp. 11–33, 2004. Journal of Medical Engineering & Technology, vol. 40, no. 4, [63] J. Guiochet, M. Machin, and H. Waeselynck, “Safety-critical pp. 199–209, 2016. advanced robots: a survey,” Robotics and Autonomous Sys- [78] H. K. Yap, F. Sebastian, C. Wiedeman, and C. H. Yeow, tems, vol. 94, pp. 43–52, 2017. “Design and characterization of low-cost fabric-based flat [64] K. Chinzei, “Safety of surgical robots and IEC 80601-2-77: the pneumatic actuators for soft assistive glove application,” in first international standard for surgical robots,” Acta Poly- 2017 International Conference on Rehabilitation Robotics technica Hungarica, vol. 16, no. 8, pp. 171–184, 2019. (ICORR), pp. 1465–1470, London, 2017. 18 Applied Bionics and Biomechanics ference on Rehabilitation Robotics (ICORR), pp. 502–507, Sin- [79] A. F. Hassanin, D. Steve, and N. M. Samia, “A novel, soft, bending actuator for use in power assist and rehabilitation gapore, 2015. exoskeletons,” in 2017 IEEE/RSJ International Conference [93] C. D. Takahashi, L. Der-Yeghiaian, V. Le, R. R. Motiwala, and on Intelligent Robots and Systems (IROS), pp. 533–538, Van- S. C. Cramer, “Robot-based hand motor therapy after stroke,” couver, BC, 2017. Brain, vol. 131, no. 2, pp. 425–437, 2008. [80] P. Polygerinos, Z. Wang, J. T. B. Overvelde et al., “Modeling [94] T. Proietti, V. Crocher, A. Roby-Brami, and N. Jarrassé, of soft fiber-reinforced bending actuators,” IEEE Transac- “Upper-limb robotic exoskeletons for neurorehabilitation: a tions on Robotics, vol. 31, no. 3, pp. 778–789, 2015. review on control strategies,” IEEE Reviews in Biomedical [81] B. Wang, K. C. Aw, M. Biglari-Abhari, and A. McDaid, Engineering, vol. 9, pp. 4–14, 2016. “Design and fabrication of a fiber-reinforced pneumatic [95] C. M. L. Hughes, P. Tommasino, A. Budhota, and bending actuator,” in 2016 IEEE International Conference D. Campolo, “Upper extremity proprioception in healthy on Advanced Intelligent Mechatronics (AIM), pp. 83–88, aging and stroke populations, and the effects of therapist- Banff, 2016. and robot-based rehabilitation therapies on proprioceptive function,” Frontiers in Human Neuroscience, vol. 9, 2015. [82] Y. Jiang, D. Chen, P. Liu et al., “Fishbone-inspired soft robotic glove for hand rehabilitation with multi-degrees-of-free- [96] M. K. Burns, D. Pei, and R. Vinjamuri, “Myoelectric control dom,” in 2018 IEEE International Conference on Soft Robotics of a soft hand exoskeleton using kinematic synergies,” (RoboSoft), pp. 394–399, Livorno, 2018. vol. 13, Tech. Rep. 6, IEEE Transactions on Biomedical Cir- cuits and Systems, 2019. [83] H. Lee, B. B. Kang, H. In, and K.-J. Cho, Design improvement of a polymer-based tendon-driven wearable robotic hand (exo- [97] H. K. Yap, A. Mao, J. C. H. Goh, and C. H. Yeow, “Design of a glove poly), vol. 16 of Wearable Robotics: Challenges and wearable FMG sensing system for user intent detection dur- Trends, , no. 1, 2017Springer, 2017. ing hand rehabilitation with a soft robotic glove,” in 2016 6th IEEE International Conference on Biomedical Robotics [84] M. Nilsson, J. Ingvast, J. Wikander, and H. von Holst, “The and Biomechatronics (BioRob), pp. 781–786, Singapore, 2016. soft extra muscle system for improving the grasping capabil- ity in neurological rehabilitation,” in 2012 IEEE-EMBS Con- [98] N. F. M. Nawi, A. H. K. Ching, N. Ibrahim et al., “Design and ference on Biomedical Engineering and Sciences, pp. 412– development approach of smart glove for post stroke rehabil- 417, Langkawi, 2012. itation,” International Journal of Engineering & Technology, vol. 7, no. 4, pp. 78–82, 2018. [85] F. L. Hammond, Y. Mengüç, and R. J. Wood, “Toward a modular soft sensor-embedded glove for human hand [99] S. Sasidharan, Smart glove: an assistive device to enhanced motion and tactile pressure measurement,” in 2014 IEEE/RSJ recovery of hand function during motor rehabilitation, [M.S. International Conference on Intelligent Robots and Systems, thesis], Arizona State University, 2015. pp. 4000–4007, Chicago, IL, 2014. [100] G. Rudd, L. Daly, V. Jovanovic, and F. Cuckov, “A low-cost [86] F. Schmitt, O. Piccin, L. Barbé, and B. Bayle, “Soft robots soft robotic hand exoskeleton for use in therapy of limited manufacturing: a review,” Frontiers in Robotics and AI, hand-motor function,” vol. 9, Tech. Rep. 18, Applied Sci- vol. 5, pp. 84–101, 2018. ences, 2019. [101] M. Haghshenas-Jaryani, C. Pande, and B. J. Muthu Wijesun- [87] P. Polygerinos, N. Correll, S. A. Morin et al., “Soft Robotics: Review of Fluid‐Driven Intrinsically Soft Devices; dara, “Soft robotic bilateral hand rehabilitation system for fine motor learning,” 2019 IEEE 16th Intertnational Confer- Manufacturing, Sensing, Control, and Applications in Human‐Robot Interaction,” Advanced Engineering Mate- ence on Rehabilitation Robotics (ICORR), pp. 337–342, Toronto, ON, 2019. rials, vol. 19, no. 12, pp. 1–22, 2017. [102] T. George Thuruthel, Y. Ansari, E. Falotico, and C. Laschi, [88] M. Haghshenas-Jaryani, C. Nothnagle, R. M. Patterson, “Control strategies for soft robotic manipulators: a survey,” N. Bugnariu, and M. B. J. Wijesundara, “Soft robotic rehabil- Soft Robotics, vol. 5, no. 2, pp. 149–163, 2018. itation exoskeleton (rehab glove) for hand therapy,” in Vol- ume 3: 19th International Conference on Advanced Vehicle [103] D. Popescu, M. Ivanescu, R. Popescu, A. Petrisor, L. C. Technologies; 14th International Conference on Design Educa- Popescu, and A. M. Bumbea, “Post-stroke hand rehabilitation tion; 10th Frontiers in Biomedical Devices, Cleveland, OH, using a wearable robotic glove,” in Smart Innovation, Systems and Technologies, pp. 259–268, Tenerife, Spain, 2016. [89] B. W. K. Ang and C. H. Yeow, “Print-it-Yourself (PIY) glove: [104] J. Nassour and F. Hamker, “Enfolded textile actuator for soft a fully 3D printed soft robotic hand rehabilitative and assis- wearable robots,” in 2019 IEEE International Conference on tive exoskeleton for stroke patients,” in 2017 IEEE/RSJ Inter- Cyborg and Bionic Systems (CBS), pp. 60–65, Munich, Ger- national Conference on Intelligent Robots and Systems (IROS), many, 2019. pp. 1219–1223, Vancouver, 2017. [105] F. Vanoglio, P. Bernocchi, C. Mulè et al., “Feasibility and effi- [90] C. G. Rose and M. K. O'Malley, “Hybrid rigid-soft hand exo- cacy of a robotic device for hand rehabilitation in hemiplegic skeleton to assist functional dexterity,” IEEE Robotics and stroke patients: a randomized pilot controlled study,” Clinical Automation Letters, vol. 4, no. 1, pp. 73–80, 2019. Rehabilitation, vol. 31, no. 3, pp. 351–360, 2017. [91] M. Li, B. He, Z. Liang et al., “An attention-controlled hand [106] H. K. Yap, N. Kamaldin, J. H. Lim, F. A. Nasrallah, J. C. H. exoskeleton for the rehabilitation of finger extension and flex- Goh, and C. H. Yeow, “A magnetic resonance compatible soft ion using a rigid-soft combined mechanism,” Frontiers in wearable robotic glove for hand rehabilitation and brain Neurorobotics, vol. 13, no. 34, pp. 1–13, 2019. imaging,” IEEE Transactions on Neural Systems and Rehabil- itation Engineering, vol. 25, no. 6, pp. 782–793, 2017. [92] B. Radder, A. I. R. Kottink, N. van der Vaart et al., “User- centred input for a wearable soft-robotic glove supporting [107] D. Popov, I. Gaponov, and J. H. Ryu, “Portable exoskeleton hand function in daily life,” in 2015 IEEE International Con- glove with soft structure for hand assistance in activities of Applied Bionics and Biomechanics 19 Communication and Electronics Systems (ICCES), pp. 341– daily living,” IEEE/ASME Transactions on Mechatronics, vol. 22, no. 2, pp. 865–875, 2017. 345, Coimbatore, India, 2017. [108] K. Shiota, S. Kokubu, T. V. J. Tarvainen et al., “Enhanced [123] K. O. Thielbar, K. M. Triandafilou, H. C. Fischer et al., Kapandji test evaluation of a soft robotic thumb rehabilitation “Benefits of using a voice and EMG-driven actuated glove device by developing a fiber-reinforced elastomer-actuator to support occupational therapy for stroke survivors,” IEEE Transactions on Neural Systems and Rehabilitation Engi- based 5-digit assist system,” Robotics and Autonomous Systems, vol. 111, pp. 20–30, 2019. neering, vol. 25, no. 3, pp. 297–305, 2017. [109] B. Kim, H. In, D. Y. Lee, and K. J. Cho, “Development and [124] K. H. L. Heung, R. K. Y. Tong, A. T. H. Lau, and Z. Li, “Robotic glove with soft-elastic composite actuators for assessment of a hand assist device: GRIPIT,” Journal of Neu- roengineering and Rehabilitation, vol. 14, no. 1, 2017. assisting activities of daily living,” Soft Robotics, vol. 6, no. 2, pp. 289–304, [110] A. L. Coffey, D. J. Leamy, and T. E. Ward, “A novel BCI- controlled pneumatic glove system for home-based neuror- [125] L. Cappello, J. T. Meyer, K. C. Galloway et al., “Assisting hand ehabilitation,” in 2014 36th Annual International Conference function after spinal cord injury with a fabric-based soft of the IEEE Engineering in Medicine and Biology Society, robotic glove,” Journal of NeuroEngineering and Rehabilita- pp. 3622–3625, Chicago, IL, 2014. tion, vol. 15, no. 1, 2018. [111] A. M. Gordon, Y. C. Hung, M. Brandao et al., “Bimanual [126] P.I.Daiya, “Daiya glove,” 2015, https://exoskeletonreport training and constraint-induced movement therapy in chil- .com/product/daiya-glove/. dren with hemiplegic cerebral palsy,” Neurorehabilitation [127] K. Toya, T. Miyagawa, and Y. Kubota, “Power-assist glove and Neural Repair, vol. 25, no. 8, pp. 692–702, 2011. operated by predicting the grasping mode,” Journal of System [112] L. Gerez, J. Chen, and M. Liarokapis, “On the development of Design and Dynamics, vol. 5, no. 1, pp. 94–108, 2011. adaptive, tendon-driven, wearable exo-gloves for grasping [128] C.-Y. Chu and R. M. Patterson, “Soft robotic devices for hand capabilities enhancement,” IEEE Robotics and Automation rehabilitation and assistance: a narrative review,” Journal of Letters, vol. 4, no. 2, pp. 422–429, 2019. NeuroEngineering and Rehabilitation, vol. 15, no. 1, pp. 9– [113] Glohera, “Hand rehabilitation glove,” 2019, http://www 14, 2018. .gloreha.com/?lang=it. [129] Z. Ouyang, P. Zhou, J. Du, and H. Chi, “Modified technology [114] D. Ryu, K.-W. Moon, H. Nam et al., “Micro hydraulic system readiness level for advanced medical devices in China,” West- using slim artificial muscles for a wearable haptic glove,” in ern Pacific Region Index Medicus, vol. 28, no. 3, pp. 217–221, 2008 IEEE/RSJ International Conference on Intelligent Robots 2015. and Systems, pp. 3028–3033, Nice, 2008. [130] ISO, “ISO 16290:2013,” 2013, https://www.iso.org/standard/ [115] Neofect, “Hand assistance glove,” 2019, http://www.neofect 56064.html. .com/en/product/rapael/. [116] J. B. Chossat, Y. Tao, V. Duchaine, and Y. L. Park, “Wearable soft artificial skin for hand motion detection with embedded microfluidic strain sensing,” in 2015 IEEE International Con- ference on Robotics and Automation (ICRA), pp. 2568–2573, Seattle, Washington, 2015. [117] B. Radder, G. B. Prange-Lasonder, A. I. R. Kottink et al., “A wearable soft-robotic glove enables hand support in ADL and rehabilitation: a feasibility study on the assistive func- tionality,” Journal of Rehabilitation and Assistive Technolo- gies Engineering, vol. 3, 2016. [118] B. Radder, G. B. Prange-Lasonder, A. I. R. Kottink et al., “The effect of a wearable soft-robotic glove on motor function and functional performance of older adults,” Assistive Technology, vol. 32, pp. 1–27, 2020. [119] J. Yi, Z. Shen, C. Song, and Z. Wang, “A soft robotic glove for hand motion assistance,” in IEEE International Conference on Real-time Computing and Robotics (RCAR), pp. 111–116, Angkor, 2016. [120] J. Yi, X. Chen, and Z. Wang, “A three-dimensional-printed soft robotic glove with enhanced ergonomics and force capa- bility,” IEEE Robotics and Automation Letters, vol. 3, no. 1, pp. 242–248, 2018. [121] Z. Yao, C. Linnenberg, A. Argubi-Wollesen, R. Weidner, and J. P. Wulfsberg, “Biomimetic design of an ultra-compact and light-weight soft muscle glove,” Production Engineering, vol. 11, no. 6, pp. 731–743, 2017. [122] S. Chatterjee, D. Hore, and A. Arora, “A wearable soft pneumatic finger glove with antagonistic actuators for finger rehabilitation,” in 2017 2nd International Conference on http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Applied Bionics and Biomechanics Hindawi Publishing Corporation

Design Criteria of Soft Exogloves for Hand Rehabilitation-Assistance Tasks

Loading next page...
 
/lp/hindawi-publishing-corporation/design-criteria-of-soft-exogloves-for-hand-rehabilitation-assistance-RKH9VWzHGN
Publisher
Hindawi Publishing Corporation
Copyright
Copyright © 2020 Juana-Mariel Dávila-Vilchis et al. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
ISSN
1176-2322
eISSN
1754-2103
DOI
10.1155/2020/2724783
Publisher site
See Article on Publisher Site

Abstract

Hindawi Applied Bionics and Biomechanics Volume 2020, Article ID 2724783, 19 pages https://doi.org/10.1155/2020/2724783 Review Article Design Criteria of Soft Exogloves for Hand Rehabilitation- Assistance Tasks 1 1 1 Juana-Mariel Dávila-Vilchis , Juan C. Ávila-Vilchis , Adriana H. Vilchis-González , 1,2 and LAZ-Avilés Faculty of Engineering, Universidad Autónoma del Estado de México, Toluca 50130, Mexico Cátedras CONACYT, Universidad Autónoma del Estado de México, Toluca 50130, Mexico Correspondence should be addressed to Adriana H. Vilchis-González; avilchisg@uaemex.mx Received 24 February 2020; Revised 3 July 2020; Accepted 7 July 2020; Published 1 August 2020 Academic Editor: Simo Saarakkala Copyright © 2020 Juana-Mariel Dávila-Vilchis et al. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. This paper establishes design criteria for soft exogloves (SEG) to be used as rehabilitation or assistance devices. This research consists in identifying, selecting, and grouping SEG features based on the analysis of 91 systems that have been proposed during the last decade. Thus, function, mobility, and usability criteria are defined and explicitly discussed to highlight SEG design guidelines. Additionally, this study provides a detailed description of each system that was analysed including application, functional task, palm design, actuation type, assistance mode, degrees of freedom (DOF), target fingers, motions, material, weight, force, pressure (only for fluids), control strategy, and assessment. Such characteristics have been reported according to specific design methodologies and operating principles. Technological trends are contemplated in this contribution with emphasis on SEG design opportunity areas. In this review, suggestions, limitations, and implications are also discussed in order to enhance future SEG developments aimed at stroke survivors or people with hand disabilities. because cartilage weakens, muscle mass decreases, and joint 1. Introduction stiffness increases [5]. More than 50 million elderly people Hand and finger motions are imperative for grasping and have difficulties to achieve accurate gripping and pinching manipulation tasks. Nonetheless, people who have suffered forces, and their range of motion (ROM) is limited as well from cerebral palsy (CP), stroke, or spinal cord injury (SCI) as their work area [6]. Therefore, people with hand disabilities can initiate a have great difficulty in accomplishing these activities of daily living (ADL) by themselves. A person with any of these prompt rehabilitation protocol in order to start recovering pathologies could present clenched fist, spasticity, uncoordi- motor skills, stop joint stiffness, and increase their indepen- nated motions, loss of strength, or diminished dexterity. dence and self-esteem [7]. Physical and occupational thera- These are consequences of a neuronal impairment that is pies are the most common treatments to recover patients’ responsible for controlling motricity, muscle endurance, movements, for example, adduction-abduction or flexion- and tonicity [1]. Worldwide, more than 15 million people extension of finger, wrist, or elbow joints. However, these are affected each year [2], and only 11.6% of the stroke survi- routines can be exhausting, time-consuming, and, relatively, vors are able to recover dexterity [3]. Patients with these dis- costly since patients require the assistance of a therapist abilities can, freely, flex their hand muscles but show whose availability is uncertain [8]. abnormal resistance when extending them [4], requiring Normally, rehabilitation programs are customized for physical rehabilitation or assistance. each patient due to their impairment, age, and anthropomet- Other hand motor deficits are caused by ageing or hand ric dimensions. Moreover, these robot procedures are classi- deformities such as rheumatoid arthritis or osteoarthritis, fied into three main assistance levels: passive assisted mode 2 Applied Bionics and Biomechanics SEG design criteria Usability (ix) Modularity Mobility (iii) Actuation (iv) Materials (x) Portability Function (xi) Customization (v) Motion (vi) Manufacture guidance (xii) Mode of (i) Rehabilitation intervention (vii) Operation (viii) Assessment & control (xiii) Costs (ii) Assistance Figure 1: Classification of soft exo-gloves design criteria. guidelines for SEG developments based on an extensive (PAM), active assisted mode (AAM), and active resistive mode (ARM) depending on the recovery status of patients review of the state of the art and of the technique from the and support of a robot [9]. last decade. Moreover, a detailed description of 91 SEG sys- Literature has reported that rehabilitation protocols can tems is provided along with implications, limitations, and be executed by robots or soft wearable devices which have suggestions for future developments. This paper is organized as follows. Section 2 presents, emerged as a therapy tool with safe human interactions, low weight, and affordable systems [10]. Particularly, SEG classifies, and discusses the criteria that are proposed for have become an alternative approach in the effort to over- SEG design based on reported devices and specific literature. come hand dysfunctions and assist patients with handling Section 3 reports SEG’s development guidelines together tasks. SEG have the ability to combine conventional therapy with the characteristics of the 91 reviewed devices. Section 4 provides a discussion concerning significant aspects (limi- with wearable systems to mimic the natural movement of fin- gers in order to increase their mobility, preventing spasticity tations, implications, and suggestions) to be taken into and joint stiffness [11]. account for future developments of SEG systems. Conclu- SEG have mainly evolved in terms of their design, fabri- sions are at the end of this document in Section 5. cation, and control [12]. Pioneering designs started using sport gloves incorporating a control system [13, 14]. Then, 2. SEG Design Criteria SEG proposals explored synthetic leather [15], rubber [16, 17], and fabrics [18, 19] to provide flexible human-robotic Hand mobility characterization in SEG designs has turned interactions as in the case of bike gloves [20]. Elastomers out to be a challenge since hand anatomy is one of the most have become the primary option to empower flexibility and complex kinematics parts of the human body with 20 DOF lightness [21]. Moreover, instead of closed palm designs for the whole of the fingers: one for abduction-adduction in (CPD) where the whole hand is covered with the glove, open every finger (thumb included); 12 for flexion-extension for palm designs (OPD) with bare hands use elastomers trying to index, middle, ring, and pinkie fingers; and three for thumb behave as a natural extension of the human hand to compete including opponent motion [29]. with skin properties in order to achieve a suitable contact In this paper, 2 function criteria, 6 mobility criteria, and 5 with objects [22, 23]. Other assistance SEG have been devel- usability criteria are proposed in order to enhance SEG oped for material handling in hazardous environments, sup- designs and enable fast developments. These design criteria port in heavy-lifting tasks [24, 25], or extravehicular tasks in are based on the aspects that have been identified from the space [26]. 91 SEG systems reported in this article and on the soft wear- Mostly, SEG systems have been driven by electrical able device’s methodology established in [28]. Figure 1 illus- energy or fluid (pneumatic or hydraulic) pressurization. trates the proposed criterion classification. Regarding electrical power supplies, tendon-driven systems Moreover, biological inspiration has come to the fore in employ linear actuators to push and pull cables embedded SEG design to emulate an animal’s motion looking for stabil- in Teflon tubes [27]. Pneumatic actuation includes fiber- ity [30] or optimal grasping tasks [31]. According to [8], SEG reinforced elastomer actuators (FREAs), inflatable chambers, should weigh less than 500 g, provide easy and comfortable or pneumatic artificial muscles, commonly known as McKib- donn-doff, and achieve 10 open-close finger cycles per min- ben muscles [28]. ute for effective actuation. Regarding SEG mechanical design, People with hand dysfunctions demand for reliable SEG authors in [29] suggest taking into account the number of to improve their quality of life. Nevertheless, the lack of joints and working DOF, the type of actuators, and the appli- affordable and accessible SEG for hand impairment patients cation. Other attributes in SEG design should adopt the char- with low-cost manufacturing processes is still a significant acteristics of a rehabilitation device which include mode of challenge. Therefore, this paper has reviewed the progress intervention (unilateral or bilateral), number of DOF, target in the field of SEG for neuromuscular rehabilitation and portion (distal, proximal, or quantity), and motion guidance assistance to overcome hand motor dysfunctions. (passive or active), among others [32]. The main contribution of this paper is the identification Based on reported literature, the next paragraphs discuss and classification of 13 design criteria to provide a set of the criteria presented in Figure 1. Applied Bionics and Biomechanics 3 SEG for assistance tasks are designed to perform three 2.1. SEG Function Criteria. SEG are classified into rehabilita- tive or assistive devices depending on their purpose [14]. SEG integral functions of the human hand: (i) finger mobilization, systems must be able to execute physical therapy and manip- (ii) holding (grasping and gripping) with high precision and strength, and (iii) manipulation for positioning and releasing ulation tasks to offer efficient and competitive devices for those with hand disabilities. Then, rehabilitation and assis- objects [8]. Assistive SEG should execute grasping, holding- tance criteria must consider the aspects discussed in the lifting, and releasing motions as continuous actions to respective paragraphs. achieve a complete manipulation [50]. To achieve stable grasping, thumb, index, and middle fingers must be included on SEG systems [35]. According to [51], soft exoglove devices 2.1.1. Rehabilitation Criterion. Rehabilitation SEG are should provide 8 N of grasping force to manipulate an object designed to help the patient regain strength, dexterity, and with a mass of 1.5 kg. coordination to recover hand functionality and range of motion (ROM) [33]. These SEG are focused on performing 2.2. SEG Mobility Criteria. From a functional perspective, specific fist motions such as full, hook, straight, and tabletop authors in [52] propose that weight, size, and power con- [18] or open-close to improve grasping tasks [34]. sumption can define an efficient soft exoglove that fits the Thumb, index, and middle finger flexion-extension is anatomical ROM of the human hand. The mass of the whole needed for strong grasping [31, 35]. Supplementary motions system should not exceed 3 kg to be considered as an assistive such as adduction-abduction are required to grasp and device [50]. These characteristics are included in criteria 3 to release objects in a more natural way [36]. Furthermore, flex- 8 (see Figure 1): actuation, materials, guidance mode, manu- ion at the interphalangeal (IP) and metacarpophalangeal facture, operation and control, and assessment that are dis- (MCP) joints with rotation at the carpometacarpal (CMC) cussed as follows. joint is necessary to reproduce thumb opposition [37]. Other SEG are able to perform wrist flexion [38], wrist radial-ulnar 2.2.1. Actuation Criterion. As aforementioned, tendon- deviation [16], or forearm pronation-supination motion [39]. driven actuators use wires to emulate human tendon func- SEG rehabilitation routines can include virtual reality in tions as flexion-extension motion. This type of actuation order to analyse the effects of brain stimulation when execut- can include Bowden cable transmissions to separate the con- ing specific tasks [40]. Patients are immersed in a game envi- trol unit from the end effector and reduce weight [53]. Also, ronment where they achieve manipulation tasks such as artificial muscle wires have been proposed to avoid friction squeezing oranges, catching butterflies, or grabbing objects [54], and shape memory alloys (SMA) have been employed [39]. Other SEG rely on neuroimaging techniques [41] or due to their elasticity [55] and high force-weight ratio [26]. provide feedback to assess a patient’s conditions and monitor On the other hand, pneumatic actuators could be embed- their progress [42]. Nevertheless, it is not enough to train the ded into inflatable air bladders [16] and into a double layer brain and do physical therapy; a successful rehabilitation sheet with curved rubber muscles [15] or made of flexible process depends on the patient’s response and their own electrostatic discharge plastic sheet materials [1, 56]. The capabilities [43]. McKibben muscles represent an affordable choice [57] and Depending on each rehabilitation protocol, the required have the ability to constrain any radial expansion during time to use a soft exoglove varies. For instance, 60 minutes pressurization [58]. Hydraulic actuators offer high load per day is recommended by [44]. Pilot tests performed by capacity [11]. [45] suggested rehabilitation sessions from 30 to 40 minutes A new trend is hybrid actuation which fits hand motion 5 days a week. Authors in [38] recommend 45 minutes but shape using soft pneumatic actuators and tendon-driven no more than 90 minutes per day to avoid SEG strain defor- operation [7], providing customization based on rigid frames mations. Authors in [46] suggest 180 minutes per week, while and soft muscles [48]. Table 1 reports the advantages and dis- authors in [39] determine that 30 minutes per day over the advantages of different SEG actuations. course of 20 sessions is necessary for a positive sizable impact When using a soft glove, patient safety must be guaran- on the impaired hand. Furthermore, to achieve a successful teed. Thus, all SEG must include different safety strategies rehabilitation program, patients should combine 30 minutes and levels in their design. For example, on cable actuation, of SEG training with 30 minutes on occupational therapy per mechanical stops, torque, or tension limiters have been day [47]. implemented [59]. Regarding pneumatic actuation, solenoid and exhaust valves are employed along with pressure regula- 2.1.2. Assistance Criterion. Eating, dressing, and writing are tors to control air flow or avoid air returns [41]. Quasistatic, everyday actions that are done unconsciously. Nevertheless, dynamic, and material failures are discussed in [60], where those tasks turn out to be a tough challenge for people with measures that can be considered in order to avoid unsafe sit- hand dysfunctions. Normally, patients depend on their fam- uations for soft robots are provided. ily or on a therapist to assist them [48]. Hence, assistive SEG Other safety levels have been applied to the electrical con- are intended to help patients to achieve manipulation tasks figuration such as emergency stops, watch dogs, or physical despite their restricted ROM, to interact with their surround- decoupling of power interfaces from logic ones by electromag- ings, and to execute ADL by themselves. These systems are netic couplings [51]. In addition, by using closed-loop control recommended when rehabilitative SEG are not enough to (CLC) schemes, sensing errors are minimized and operation overcome patient stiffness [49]. in a stable regime is ensured to avoid hyperextension at the 4 Applied Bionics and Biomechanics Table 1: Actuators for SEG systems. Actuation Types Advantages Disadvantages (i) Cable paths reduce friction (i) Complex transmissions (i) Muscle wires (ii) Provides continuous force (ii) Continuous hours of operation are restricted Electrical (ii) Tendon-driven (iii) Stores energy (iii) Nonlinearity of the system makes control (iii) Shape memory alloys (iv) Commercial availability difficult (i) Requires compressed air (i) FREA (i) Allows multiple DOF (ii) Requires a reservoir (ii) Supports their structural shape (ii) Inflatable chambers Pneumatic (iii) Inaccurate forces (iii) Pneumatic artificial (iii) Allows adaptability (iv) Problems with leaks muscles (iv) Lightweight (v) Portability is restricted (i) High load capacity and power (i) Heavy systems supply (ii) Problems with leaks Hydraulic (i) Fluid chambers (ii) Low cost (iii) Portability is restricted (iii) Allows multiple DOF (iv) Requires a reservoir and a pump wrist or overflexed fingers, for instance [20]. At the program- grip strength and reduce pressure on the skin [51, 71]. Addi- ming level, haptic feedback is also included to prevent acci- tionally, silicon materials offer stable fastening and prevent dents [61]. slippage [72]. These synthetic polymers are easy to wash More specialized safety strategies related to robots can be and do not absorb sweat compared to textile materials [23]. considered, such as safety standards or means to guarantee SEG made of fabrics have low cost and offer minimal mechanical impedance to finger motion [73]. Hence, coated system dependability [62] as fault prevention, fault removal, fault forecasting, and fault tolerance [63]. Being safety a pri- fabric SEG systems with thermoplastic polyurethane (TPU) ority aspect, it constitutes a current research area by itself and actuators are recommended for customization and to avoid must be taken into account in the development of SEG sys- slipping or muscle expansion problems [74]. tems. Concerning rehabilitation robots, ISO-IEC 80601-2- Actuators made of fabrics work at lower pressures than elastomer actuators due to their inherent stiffness [75]. 78 must be taken into account. Many specialized documents are recommended for readers interested in this topic and for Therefore, several researchers have work on design, charac- researchers and engineers working in SEG design (see, for terization, manufacture, and evaluation of soft elastomer instance, [64–66]). actuators for hand [76–78] and wrist [79] rehabilitation. Additionally, relevant features for actuators have been To match and support finger flexion-extension, some designs include multisegment elastomers with fiber rein- identified in SEG literature or proposed in this paper. For instance, current developments have focused on improving forcement [80, 81] or corrugated fabric layers [41, 43] which actuator design to tackle more DOF [67]. During SEG assem- are pressurized from 70 kPa up to 375 kPa [75]. Other bly, the actuators are mounted into the dorsal side of the designs include rigid plastic hoops [67] or nylon strings hand to avoid finger movement obstruction [68] and can be [82] to avoid radial deformations in FREA. Material selection has also played a significant role in fas- removed from the glove [69]. Actuators must not affect the active ROM of the finger joints and should allow free motions tening the actuators to the glove or fingers in a safe way. with more contact area for grasping tasks in a compliant Mostly, SEG proposals have employed magnets [83] or straps manner [21]. made of Velcro® [8, 18], fabrics [84], and rubber [24]. Other Furthermore, actuators should take less than 4 s for full designs had opted for sewing the components [71] or sepa- rating the system from the actuators to reduce weight. Actu- grasping [1]. The length of actuators should not be longer than the length of the fingers to avoid mismatching problems ators can be attached to the wrist through elastomer bracelets between them [23]. Actuators with low power consumption [39] or synthetic hide covers [25, 31]. and continuous hours of operation are recommended. 2.2.3. Manufacture Criterion. Mobility is also determined 2.2.2. Material Criterion. To enhance SEG operation, by manufacturing processes since specific elements can researchers continue to seek compliant, flexible, and light- be obtained by particular methods that, additionally, can weight materials to easily conform hand-finger anatomy with determine the weight and dexterity of the system. Conven- the shape of an object [41]. Hence, the payload capacity of tional manufacturing procedures involve polymer casting molds [85], reinforcements and inclusions [11], additive elastomers has been exploited to obtain an elastic modulus similar to that of human tissues and avoid cumbersome manufacturing, thin-film manufacturing, shape deposition designs [70]. manufacturing, and bonding [86]. Nonferromagnetic materials such as nylon, neoprene, Mostly, 3D printing two-part mold has been employed for polyester, or synthetic leather have been selected as compli- SEG spacers [23], cable guides [73], and elastomer actuators [87] where one mold is used to create a fluid chamber inside ant and affordable options to increase conformability and Applied Bionics and Biomechanics 5 Assistance Control mode DOF Number of motion Palm Fingers design Tasks Actuation Application 40 Material Assessment Function Force Pressure Weight Figure 2: Frequency of SEG aspects reported in Table 3. the actuators and the second one is addressed to create a fabric adduction-abduction [85] and perform flexion-extension layer on top of the actuators [41, 43]. Nevertheless, low repeat- motions [39, 71]. More sophisticated SEG systems have already begun an age that allows patients to perform a ability is the main drawback during this process [48]. Recent developments involve thermomethods [34], desired movement. When patients are able to achieve func- inverse flow injection [42, 82], lost wax molding [88], or tional ROM, the system will have no effect on the hand fused deposition modeling with 3D printing at home to [41] or will create an opposite force to improve the power reduce SEG costs and facilitate its acquisition [89]. However, of the patient. there is still room to improve SEG materials and fabrication Most of the reported SEG systems focus on PAM, a few with low costs. on AAM as well as on the combination of active and passive New trends are oriented to hybrid designs where they modes (see Figure 2). combine rigid and soft components to obtain more hand poses and more DOF [90] and provide active training that 2.2.5. Operation and Control Criterion. SEG operation is encourages user participation [91]. defined by their type of actuation and their components. Tendon-driven wires require servomotors, gearboxes, spools, 2.2.4. Motion Guidance Criterion. SEG are designed to follow and force/torque sensors to move them. Pneumatic systems specific trajectories defined by a therapist depending on the require a compressor, electrovalves or proportional valves, impairment of the patient. These trajectories seek to achieve pressure sensors, or regulators. All these components are a functional ROM during both active and passive modes. controlled on a data acquisition board which is plugged to a SEG is aimed at promoting active finger flexion and pas- PC or uses Bluetooth as a communication interface for the sive extension to increase patient autonomy during eating or SEG system [74]. drinking tasks [46, 92]. In the active assistance mode, Different schemes have been proposed to operate and patients attempt to move their hand and SEG are an addi- control SEG systems; for example, in [14], Faulhaber 1226 tional aid to complete the desired ROM [93] whereas in the 006B motors, CompactRIO board, and LabVIEW® are used. passive assistance mode, the exoglove provides all the assis- Authors in [21] use DCX22 motors, a control board tance to guide the desired movement [94]. In the patients’ TMS320F2808®, and Simulink®. Additionally, graphical user force recovery processes, effective SEG systems should, actively, participate with intensive training based on active interfaces (GUI)® have been implemented as a communica- tion channel for SEG systems [88, 91]. A broad range of and repetitive practical motions [95]. operation and control possibilities exists to select microcon- SEG should combine active and passive mobilizations for successful hand rehabilitation. For example, authors in [15] trollers and interfaces relying on desired real-time response, accuracy, number of components involved in the operation provide active extension on each finger. In [8], SEG also exert passive extension with active flexion and thumb opposition and control strategies, and specific requirements of each for grasping tasks. Other systems include active finger SEG system. Number of soft exogloves RT AT AT-RT S troke Hand disabilities Other Grasping Pinching Both Closed Open Tendon-driven Pneumatic Hydraulic Passive Active Active-passive < 10 > 10 All Ind-Mid-Thu All except Thu Elastomer Fabrics Other < 100 g 100-200 g > 200 g < 10 N 10-20 N > 20 N < 100 kPa 100-300 kPa > 300 kPa CLC OLC Forces ROM Other 6 Applied Bionics and Biomechanics to move specific hand joints that are connected to the afore- Normally, open-loop control (OLC) and closed-loop control strategies are implemented during SEG operation. mentioned muscles. Due to stable behaviours, force myogra- OLC schemes have used springs [1, 34] or mechanical phy (FMG) signals have been proposed to control the intention of the movement on SEG systems [20]. switches [54] for manual operation where patients are able to drive an actuator to accomplish a specific task [96]. OLC Motor impairment scales are applied to evaluate patient strategies require the system to be stable by construction. ROM to determine SEG operation ranges before starting an To regulate the desired variables or to track specific trajecto- aided rehabilitation process. These scales serve for the eval- ries that ensure patient safety while using a soft exoglove, uation of the damage that each patient has. According to [38, 105], patients with an Ashworth spasticity index (ASI) CLC strategies are implemented [97]. To achieve acceptable motions in CLC schemes, sensors are directly attached to value less than or equal to three can use a soft exoglove. A SEG [98, 99] without the patient worrying about making modified Ashworth scale (MAS) value less than or equal to accurate movements. two defined the use of a soft exoglove for active flexion- Proportional (P) [68], proportional-derivative (PD) [100, extension, according to [106]. SEG operation is also based on the functional independence measure (FIM) of the 101], or proportional-integral-derivative (PID) [15, 71] con- trollers are widely implemented for flow and force regulation. patient by which the value goes from 1 to 7 depending on Pulse width modulation (PWM) signals have been used to the assistance intensity [45]. Thus, for values above 3, open and close solenoid valves [51] and can be implemented patients present more autonomy [36]. in many control strategies for different applications. Other 2.2.6. Assessment Criterion. To ensure patient safety and SEG kinds of controllers can be used depending on the system nature and on the task objective. For instance, nonlinear con- operation, several tools such as joint contractions [31, 54], trollers, fuzzy approaches, or optimal linear control schemes bending angles [71], 3D visual motion analysis [11], or opti- could be developed for specific SEG systems. For instance, cal ROM at specific joints [44] have been employed to evalu- [102] provides an interesting review of soft robotic manipu- ate SEG performance. Other methods have opted for using mathe lator control strategies that could be considered to be applied matical models together with the finite element in SEG systems. method (FEM) for hand and finger trajectory characteriza- SEG operation is based on force and position require- tion [67]. To assess patient satisfaction when using SEG sys- ments to emulate human hand functions. These require- tems, questionnaires have been considered [88]. ments, among others, are taken into account to define the SEG assessment can be also done based on the blocked control strategy to be synthesized. For example, SEG should [48], grasping [21], pinching [44], or fingertip [14] forces that have less than 10 minutes of setting time to become a useful are quantified using bottles, cups, balls, telephones, cans, or tool for therapists [103]. Regarding fluid actuation, 10 N to fruits with variable mass, size, and texture [44, 51]. For cylin- 15 N are required for grasping tasks [11, 41]. SEG must be drical objects, the diameters go from 50 mm to 120 mm [21, able to generate 7 N per finger or around 25 N on the whole 75] with a mass of 300 g [107]. Experimental tests on SEG hand with distributed forces along the fingers to minimize assessment have been carried out with dummy hands [71] pressure location points, according to [34]. Normally, actua- and healthy individuals [44] or combining healthy people tors with variable stiffness require 120 kPa for pinching and and stroke survivors [75]. Other SEG evaluations perform 160 kPa for grasping [18] while multisegment actuators tasks with/without a soft exoglove and compare them [46, require between 345 kPa and 400 kPa for flexion motion 92]. ROM data have been collected when using a soft exo- [51]. Desired joint ROM define positions to be reached by glove and without it [31]. the patient when using a SEG system and provide reference To assess hand function and ROM using SEG systems, variables to be controlled. patients undergo coordination and dexterity tests. For To evaluate SEG effectiveness in rehabilitation or assis- example, the Kapandji score is used to evaluate thumb tance tasks, surface electromyography (EMG) has been performance on pinching and grasping tasks [108]. SEG implemented to detect user movement intentions [53], point assessment also considers the motricity index test (MIT) out muscle contractions [16], control finger motion, and [105], the Fugl-Meyer assessment (FMA) [46], the nine- force level activation of muscles [90] since this is a noninva- hole peg test (NHPT) [38], the Jebsen-Taylor hand test sive procedure that prevents muscle injuries. (JTT) [44], the box and block test (BBT) [11], the Purdue During gripping tasks for finger flexion-extension, EMG pegboard test (PPT) [45], or some writing tasks [109]. signals are captured from the extensor digitorum communis For each patient, one or more of the aforementioned (EDC) muscle together with the flexor digitorum superficialis methods could be chosen by his/her motor impairment or (FDS) [20, 50] or with the flexor digitorum profundis (FDP) by the therapist in charge of the respective rehabilitation pro- muscle [43, 73] since these muscles have been used and tested tocol in order to assess SEG systems. to work properly when implementing EMG procedures and Some authors have focused more on statistical analysis due to the number of fingers they are connected with. Then, about user condition than SEG performance [5, 40]. They data obtained from a set of electrodes are amplified, filtered, seek for a specific target group, rehabilitation time, training quantified, and converted from analog to digital signals dur- tools, age, or gender, for instance. ing SEG use [104]. This electrical stimulation should be mon- itored at least every 10 minutes to avoid muscle fatigue [103]. 2.3. SEG Usability Criteria. To guarantee a friendly and com- EMG signals can be used as control inputs when it is required fortable SEG use, modularity, portability, customization, Applied Bionics and Biomechanics 7 more beneficial than unilateral mode since patients can inte- mode of intervention, and cost criteria must be considered to develop a soft exoglove with particular characteristics as easy grate healthy and paretic hand motions during rehabilitation to put on and operate, working in an intuitive way, and hav- therapy [75]. The bilateral mode is supported by a master- slave therapy concept where healthy limbs act as masters ing low cost. These criteria are discussed below. and soft devices as slaves [101]. Then, healthy limbs become 2.3.1. Modularity Criterion. SEG designs have opted for mod- a support for paretic limbs whereas devices working in the ular configurations to ease donn and doff as in the cases of unilateral mode only exercise the impaired limb [111]. Bilat- [21, 72, 83]. Connections can be assembled to work on tar- eral mode rehabilitation could be recommended by the ther- geted tasks, and actuators are mounted one by one [39]. apist as a function of the impairment. Then, SEG design Besides, modular designs for bending motions with deploy- could consider the mode of intervention depending on the able mechanisms have been adopted to reduce weight and associated rehabilitation protocol. allow natural motion [48]. SEG quality can be improved by a modularized system with relatively low cost customization, 2.3.5. Cost Criterion. It has been noted that researches are easy maintenance, and low power consumption [23]. Addi- more interested in the functionality of their products than tionally, modular architectures allow for the replacement of in their price, since only few works report SEG costs. How- feasible SEG components [89]. Based on this information, ever, SEG cost will determine one of the aspects for the suc- modularity is highly recommended as one of the main char- cess of an exoglove as a commercial product. Therefore, acteristics of SEG systems. designers could generate low-cost readily available SEG sys- tems. For instance, authors in [34] propose that the assembly 2.3.2. Portability Criterion. To cope with patients’ demands should cost less than $30 USD in order to be a competitive and to guarantee continuous rehabilitation protocols, the choice. Another proposal establishes that manufactuing and use of SEG outside clinics has become a main design con- electronics should be less than $200 USD [100]. According cern to foster external rehabilitation [38, 110]. Nevertheless, to [52], soft exosuits for the upper limb should cost less than to achieve this objective, SEG performance depends on the $1000 USD, $465 USD for the elbow, and $470 USD for the number of hours they can operate continuously without hand. A detailed description about the component cost of having a fixed power supply. According to [51], an effective these configurations could be found in [59]. SEG costs can soft exoglove should achieve, without problems, 2 hours of vary due to the type of actuation, the type of components continuous operation or from 4 to 6 hours of intermittent and materials, the weight, and the country where they were operation. developed [112]. Moreover, the runtime of batteries should be more than Remarkable results about cost analysis between conven- one hour in order to guarantee the development of a rehabil- tional and aided therapy show that SEG rehabilitation is itation protocol session [100] until its completion or exert more affordable than therapist assistance since the reported from 15 to 20 minutes of passive guidance [88]. Normally, cost associated with aided therapy is almost three times less lithium-polymer batteries are used since they can last 3.8 expensive than the conventional one [45]. hours of continuous operation [23, 51]. Currently, Neofect™, Glohera™, and Bioservo™ compa- Patients should take physical therapy sessions at reha- nies have already patented their systems which have been bilitation facilities as well as at home [46, 92] in order to commercially exploited for hand rehabilitation and assis- perform exercises on their own and not only depend on tance. However, these commercial systems are available only the availability of therapists [34]. SEG must be lightweight in some countries and are, relatively, expensive. Importation to allow their transportation [31, 73]. Thus, control unit and shipping costs must be added to final prices for countries boxes should be set up independently of the glove to min- and locations where these systems are not available. imize additional load [74]. Some power supply designs include waist belts [43, 51, 89], backpacks [73], boxes 3. SEG Design Guidelines [11, 50], vests [84], waist pockets [53, 59], pockets [44], or a separate section located on another part of the human Based on the information provided in Section 2, Table 2 sum- body [25, 31]. marizes some of the main aspects related to the 13 proposed design criteria for SEG developments. 2.3.3. Customization Criterion. As established by [18], con- At present, SEG approaches are focused on improving formability, adaptability, and customization are some fea- functionality, strength, DOF, and ROM for object manip- tures that can be taken into account to fit, properly, the ulation. Figure 2 and Table 3 provide information for each hand of a patient and generate a compliant soft exoglove. of the 91 SEG systems reviewed in this paper, associated Particularly, customization affects SEG operation since each with the following 15 aspects: (1) function: robot rehabili- finger length varies due to sex, age, and finger palm size tation (RT) or assistance tasks (AT); (2) application: hand [17]. Thus, fasteners [71] and Velcro® straps [8, 110] have disability, stroke survivors, or SCI; (3) task: grasping, been used to attach, conveniently, SEG to hands. Otherwise, pinching, or manipulation; (4) palm design: OPD or deviations from a nonappropriate size or form may restrict CPD; (5) type and number of actuators: tendon-driven, hand movement or cause discomfort during SEG use [21]. pneumatic, or hydraulic; (6) assistance mode: PAM or 2.3.4. Mode of Intervention Criterion. To increase hand func- AAM; (7) DOF per finger; (8) targeted fingers; (9) motions: tion rehabilitation, a bilateral mode in SEG systems results flexion-extension, adduction-abduction, opponent, ulnar/ 8 Applied Bionics and Biomechanics (6) AAM should be the priority motion guidance for Table 2: Proposed criteria and considerations for SEG system design. SEG rehabilitation Type Criteria Considerations (7) Mostly SEG provide more than 10 DOF to reach hand motor function Stroke survivors and Rehabilitation hand disabilities (8) A complete hand characterization must be included Function Grasp, grab, pinch, lift, to tackle more DOF Assistance hold, and release tasks (9) All SEG provide, at least, flexion-extension motion. Cable-driven, pneumatic, Actuation Furthermore, adduction-abduction and opponent or hydraulic motions are desirable Fabric, synthetic leather, Materials neoprene (10) Elastomers have become the main material choice due to their flexibility, lightness, and adaptability Motion guidance Passive or active Mobility Manufacture 3D printing+other procedures (11) SEG systems should have a total mass of less than Operation & control 45 minutes/day, OLC or CLC 200 g to enhance their efficiency 7N/finger or 10 to 15 N Assessment (12) SEG should provide, at least, 5 N per finger to exe- for 1 kg objects cute most of ADL Open or closed palm Modularity (13) Regarding pneumatic actuation, SEG should work configuration between 100 and 300 kPa 500 g or less than 3 kg for Portability the whole system (14) CLC controllers are preferable to OLC in order to Usability Right size, fasteners ensure patient safety and system precision. Particu- Customization or Velcro® straps larly, PD controllers have been mostly implemented Mode of intervention Bilateral or unilateral (15) Fingertip forces, ROM, and EMG are the most used Costs Assembly less than $30 USD tools to evaluate SEG effectiveness Table 3 provides detailed information related to the 15 radial deviations, and pronation-supination; (10) material; aspects illustrated in Figure 2 for 91 devices that have been (11) weight; (12) force; (13) pressure; (14) control: CLC or analysed in order to identify, classify, and discuss the 13 OLC; and (15) assessment. aforementioned criteria and to establish the previous 15 Figure 2 provides information related to the number of design guidelines for SEG systems. For example, the third soft exogloves that have been developed in the last decade, system has eight DOF, focuses on grasping assistance tasks, being characterized by particular aspects. For example, the has a closed palm configuration, is passively driven (CLC) most important number of SEG systems that have been by cables, and performs flexion/extension of 3 fingers. developed is focused on the passive assistance mode, CLC From the previous information reported in this paper, predominate over open-loop strategies, elastomers are pre- five core SEG developers have been identified and have ferred to other types of material, hydraulic actuation is not marked trends in the design of soft exoglove systems. Hong significant compared to the number of SEG devices using Kai Yap is the author with the highest number of SEG contri- tendon-driven or pneumatic actuation, and SEG have been butions (see Table 3, items 25-31). developed, mainly, to cope with stroke and hand disabilities The number of SEG developments, from the last ten as well as with rehabilitation and assistance problems. years, is plotted in Figure 3. According to literature, 2017 Based on what has been presented so far, the following was the most productive year with 21 of the 91 contributions SEG design guidelines are highlighted in order to be consid- reported in this paper. ered when developing new SEG systems. 4. Discussion (1) Rehabilitation and assistance tasks should be included in a single soft exoglove In order to provide technical solutions for hand rehabilita- tion and assistance, multiple endeavours have been done (2) SEG are primarily designed for stroke survivors and during the last three years about SEG developments [128]. people with hand disabilities This review has identified areas of opportunity for the (3) Grasping is the main assistance task that has been improvement of soft exogloves that are used in aided rehabil- addressed by SEG systems itation protocols and assistance tasks. Four main circum- stances have motivated researchers to satisfy popular (4) SEG have been diversified for both OPD and CPD demand and increase SEG development since they represent depending on the actuation an alternative and affordable approach to overcome hand (5) Tendon-driven and pneumatic are preferable types disabilities. These circumstances are related to the increase of actuators in the number of people with hand motor deficits, to poor Applied Bionics and Biomechanics 9 Table 3: Soft exoglove design criterion classification. Palm Assistance Weight Force Pressure SEG # Function Application Task Actuation (number) DOF/finger Fingers Motion Material Control Assessment design mode (g) (N) (kPa) (Ref.) 1 Stroke survivors, Index, middle, Fingertip and AT Grasping CPD Tendon-driven (1) PAM 8 Flexion-extension Synthetic latex 80 18 — CLC [14] SCI thumb blocked force 2 Index, middle, AT-RT Hand disability Grasping CPD Tendon-driven (3) PAM 8 Flexion-extension Synthetic latex —— — CLC (PI, PD) EMG [17] thumb 3 Stroke survivors, Index, middle, AT Grasping CPD Tendon-driven (3) PAM 8 Flexion-extension Synthetic latex 194 20 — CLC EMG [19] SCI thumb 4 Index, middle, Silicone AT Hand disability Grasping OPD Tendon-driven (2) PAM 8 Flexion-extension — 20 —— EMG [21] thumb KE-1300 5 Grasping, Index, middle, Flexion-extension, Silicone AT Hand disability OPD Pneumatic (4) PAM 11 350 22.5 300 CLC ROM, grip strength [23] pinching thumb adduction-abduction KE-1300 T 6 Grasping, Index, middle, Flexion-extension, Silicone AT SCI OPD Tendon-driven (3) PAM 9 104 10.3 — CLC Grip strength [22] pinching thumb opponent KE-1300 T 7 Grasping, Pneumatic All except RT Hand disability OPD PAM 14 Flexion-extension, curl Neoprene 160 —— CLC FEM analysis [8] pinching elastomer (4) thumb 8 Grasping, Hydraulic Flexion-extension, AT-RT Hand disability OPD PAM 15 All Neoprene —— 400 CLC ROM, EMG [51] pinching elastomer (2) opponent 9 Hydraulic Flexion-extension, AT-RT SCI Grasping OPD PAM 15 All Textile — 10-15 — CLC ROM, BBT [11] elastomer (2) opponent 10 Grasping, Hydraulic Flexion-extension, AT Hand disability OPD PAM 15 All Elastomer — 14 413 CLC ROM, EMG [50] holding elastomers opponent Flexion-extension, RT Stroke survivors Gripping OPD Pneumatic PAM 16 All opponent, —— — — — ROM: MI, BBT, FIM [36] adduction-abduction Flexion-extension, 12 Hemiplegic Gripping, AT-RT OPD Pneumatic PAM 16 All opponent, —— — — — ROM: MI, NHPT test [105] patients pinching adduction-abduction Flexion-extension, 13 Gripping, AT-RT Stroke, SCI OPD Pneumatic AAM-PAM 16 All opponent, —— — — — ROM: NHPT, FIM test [45] pinching adduction-abduction Flexion-extension, 14 Gripping, ROM: MI, RT Stroke, SCI OPD Pneumatic AAM-PAM 15 All opponent, —— — — — [113] pinching NHPT test adduction-abduction Flexion-extension, RT Stroke survivors Gripping OPD Electrical (5) PAM 15 All, wrist opponent, —— — — — Ashworth index [38] adduction-abduction AT — Grasping CPD Hydraulic PAM 14 All Flexion-extension — 2620 12 550 CLC Pressure regulating [114] Flexion-extension, 17 Stroke survivors, Grasping, All, wrist, ROM: FMA, OPD Tendon-driven AAM-PAM 14 opponent, —— — — — RT [39] SCI manipulation forearm JTT, PPT pronation-supination Flexion-extension, 18 Grasping, All, wrist, RT Stroke survivors OPD Tendon-driven AAM 14 opponent, —— — — — — [115] manipulation forearm pronation-supination Flexion-extension, 19 Stroke survivors, Grasping, All, wrist, radial-ulnar RT OPD Tendon-driven AAM-PAM 14 Elastomer 132 —— CLC ROM [47] SCI manipulation forearm deviations, pronation- supination 20 Stroke survivors, AT-RT Grasping, pinching CPD Tendon-driven (5) AAM-PAM 12 All Flexion-extension Lycra, fabrics — 15 — CLC Grip force [73] SCI 21 Middle, ring, AT Hand disability Grasping, pinching CPD Tendon-driven (3) PAM 8 Flexion-extension Synthetic leather 700 20 — CLC (PID) Grasping power test [84] thumb 22 Flexion-extension, AT-RT Hand disability Grasping OPD Tendon-driven (5) PAM 15 All Silicone rubber — 17.25 165 CLC ROM: tactile pressure [85] adduction-abduction 10 Applied Bionics and Biomechanics Table 3: Continued. Palm Assistance Weight Force Pressure SEG # Function Application Task Actuation (number) DOF/finger Fingers Motion Material Control Assessment design mode (g) (N) (kPa) (Ref.) 23 Flexion-extension, RT Hand disability Grasping OPD Tendon-driven (5) PAM 15 All Elastomer —— — — — [116] adduction-abduction 24 Material AT Grasping CPD Tendon-driven PAM — All — Fabrics —— — CLC Lift forces [24] handling AT-RT Hand disability Grasping, pinching CPD Pneumatic (5) AAM 14 All Flexion-extension Fabrics 200 — 160 CLC ROM [18] 26 All except RT Stroke survivors Grasping CPD Pneumatic (4) PAM 12 Flexion-extension Fabrics 200 9.25 200 CLC fMRI, optical ROM [41] thumb 27 Grasp Flexion-extension, AT Grasping, releasing CPD Pneumatic (5) PAM 15 All Lycra, fabrics 170 13.6 153 CLC Optical ROM, EMG [43] pathologies opponent AT-RT Stroke survivors Grasping, releasing OPD Pneumatic (5) AAM 14 All Flexion-extension Neoprene 150 — 100 CLC ROM, torque [1] 29 All except AT-RT Stroke survivors Grasping, pinching OPD Pneumatic (4) PAM 12 Flexion-extension Lycra, fabrics 180 10.2 120 CLC fMRI, optical ROM [106] thumb 30 Grasping, lifting, AT Stroke survivors OPD Pneumatic (5) AAM-PAM 14 All Flexion-extension Lycra, fabrics 180 12-36 120 CLC ROM, EMG [44] releasing 31 Grasping, AT-RT Hand disability CPD Pneumatic (5) AAM 14 All Flexion-extension Fabrics 99 13.6 275-375 CLC Gripping force [75] manipulation 32 Index, middle, AT Hand disability Grasping CPD Tendon-driven (1) PAM 8 Flexion-extension Synthetic latex 500 10 — CLC ROM, gripping force [53] thumb 33 Index, middle, AT Muscle weakness Grasping CPD Tendon-driven (2) PAM 8 Flexion-extension Neoprene 1200 10 — CLC (PD) ROM, gripping force [59] thumb, elbow 34 Index, middle, AT Hand disability Grasping CPD Tendon-driven (3) PAM 8 Flexion-extension Neoprene 500 —— CLC (PD) ROM, gripping force [52] thumb AT-RT Heavy tasks Grasping holding CPD Tendon-driven (3) PAM 14 All Flexion-extension Fabrics 770 —— CLC Grasping force [25] 36 Grasping, Pneumatic, All except AT-RT Hand disability OPD PAM 12 Flexion-extension Nylon 150 2.5 230 CLC Bending forces, EMG [30] manipulation hybrid (4) thumb 37 Grasping, Pneumatic, RT Hand disability CPD PAM 14 All Flexion-extension —— — 165.4 CLC ROM, joint angles [48] manipulation hybrid (5) 38 Hand Pneumatic RT Grasping OPD PAM 14 All Flexion-extension Textile —— 526 CLC ROM, fatigue test [42] pathologies elastomers (5) 39 Flexion-extension, AT Hand disability Grasping, pinching CPD Pneumatic (5) PAM 15 All Synthetic leather 135 9 200 CLC ROM, EMG [16] opponent Grasping, RT Stroke survivors releasing, CPD Pneumatic (5) PAM 14 All Flexion-extension Lycra — 15 60 CLC (PID) ROM: FMA, BBT [15] pinching 41 Index, middle, RT Hand disability Grasping CPD Tendon-driven (3) PAM 8 Flexion-extension Synthetic leather —— — CLC ROM, joint angles [31] thumb 42 Index, middle, AT-RT Hand paralysis Grasping CPD Tendon-driven (1) PAM 8 Flexion-extension Polyester fiber 50 35 — CLC ROM, EMG [35] thumb 43 Index, middle, AT-RT Hand disability Gripping CPD Tendon-driven (3) PAM 8 Flexion-extension Fabrics —— — CLC ROM, gripping force [117] thumb 44 Index, middle, AT Older adults Gripping CPD Tendon-driven (3) PAM 8 Flexion-extension Fabrics 85 —— CLC Pinch strength, JTHFT [118] thumb 45 Pinching force, AT-RT Hand disability Gripping CPD Tendon-driven (5) PAM 14 All Flexion-extension —— — — CLC [46] JTHFT test RT Hand disability Grasping CPD Tendon-driven (5) PAM 14 All Flexion-extension Fabrics —— — CLC ROM, joint angles [54] Flexion-extension, RT Stroke survivors Grasping OPD Pneumatic (5) PAM 16 All opponent, Nylon — 20 — CLC (PID) ROM, joint angles [71] adduction-abduction Applied Bionics and Biomechanics 11 Table 3: Continued. Palm Assistance Weight Force Pressure SEG # Function Application Task Actuation (number) DOF/finger Fingers Motion Material Control Assessment design mode (g) (N) (kPa) (Ref.) 48 Spring RT Stroke survivors Grasping CPD PAM 14 All Flexion-extension Synthetic leather 200 22.59 —— Bending force [34] mechanism (5) RT Stroke survivors Grasping OPD Pneumatic (5) PAM 14 All Flexion-extension —— — — CLC Electroencephalography [110] EMG, ROM, AT-RT Stroke survivors Gripping OPD Pneumatic (5) PAM 14 All Flexion-extension Elastomer — 41.8 200 CLC [89] gripping force 51 Grasping, AT Hand disability CPD Cable-driven (4) PAM 11 All except little Flexion-extension — 250 16 — CLC ROM, pinching force [107] manipulation 52 All except RT Hand disability Grasping CPD Pneumatic (4) PAM 12 Flexion-extension Nylon — 3 200 — FEM & ROM [67] thumb 53 Grasping, Pneumatic artificial AT Hand disability OPD PAM 14 All Flexion-extension Fabrics 161 10 200 — FEM, fingertip force [119] manipulation muscles (5) 54 Grasping, TPU, Pressure regulation AT-RT Older adults OPD Tendon-driven (5) PAM 14 All Flexion-extension 50 40 60 CLC [120] manipulation NINJAFLEX™ & fingertip force 55 22 pinch, 48 Pinching and AT-RT Hand disability Grasping, pinching OPD Tendon-driven (5) PAM 14 All Flexion-extension TPU 330 — CLC [74] grasp grasping forces 56 Shape memory ROM, fingerti p-tendon AT-RT Hand disability Grasping CPD PAM 14 All Flexion-extension Fabrics — 40 — CLC [55] alloys (5) force 57 Grasping, All except AT Stroke survivors OCP Tendon-driven (5) PAM 12 Flexion-extension Polymer 340 —— CLC Gripping force [4] manipulation thumb 58 CP, stroke Thumb, index, Grasping and AT Grasping CPD Tendon-driven (3) PAM 8 Flexion-extension Synthetic leather 55 48 — CLC [112] survivors and middle fingertip forces Flexion-extension, 59 Shape memory radial abduction, AT Supportive aid Grasping CPD PAM 12 All except little Synthetic leather 85.03 11 — CLC Grasping force, ROM [121] alloys (5) palmar abduction, opposition RT Stroke survivors Grasping OPD Pneumatic (1) PAM 3 Index Flexion-extension Ecoflex™ 00-30 — 1.17 30 CLC Bending angle [122] 61 Older adults, Pneumatic artificial Grasping and AT Grasping, pinching CPD PAM 14 All Flexion-extension Rubber — 5.7-14, 20-25 500 CLC [13] hand disability muscles (5) pinching forces 62 Thumb, index, Writing tasks, AT SCI Grasping CPD Tendon-driven (3) PAM 8 Flexion-extension —— 7.39 — CLC [109] and middle grasping force 63 Pneumatic, AT-RT Stroke survivors Grasping CPD PAM 14 All Flexion-extension Nylon — 5 300 CLC Ashworth test [7] tendon-driven (5) 64 All except RT Stroke survivors Grasping CPD Tendon-driven (4) PAM 12 Flexion-extension —— — — CLC FMA test [123] thumb 65 Pneumatic RT Hand disability Grasping OPD PAM 14 All Flexion-extension Nylon 280 11.27 250 CLC Grasping test FEM [82] FREAs (5) 66 Pneumatic RT Hand disability Grasping OPD PAM 1 Thumb Opposition Elastomer 586 — 150 CLC Kapandji test [108] FREAs (1) 67 Pneumatic AT Stroke survivors Grasping OPD PAM 14 All Flexion-extension Silicone rubber 207 — 200 CLC Bending force FEM [124] FREAs (5) 68 Grasping, AT SCI OPD Pneumatic (5) PAM 14 All Flexion-extension Fabric 77 15 172 CLC Lifting force [125] manipulation AT Stroke survivors Grasping CPD Electrical PAM 15 All, wrist Flexion-extension Neoprene fabric —— — CLC BBT test [99] RT Stroke survivors Grasping CPD Tendon-driven (1) PAM 3 Index, wrist Flexion-extension Lycra —— — CLC ROM [98] 71 Kinesthetics, Index and Virtual reality RT Pressing OPD Pneumatic (2) PAM 6 Flexion Silicone rubber — 16.66 210 CLC [61] haptic feedbacks middle haptic feedback 72 Grasping, Flexion-extension, AT Hand disabilities CPD Pneumatic (4) PAM 15 All — 65 —— CLC Grabbing force [126] holding opposition 12 Applied Bionics and Biomechanics Table 3: Continued. Palm Assistance Weight Force Pressure SEG # Function Application Task Actuation (number) DOF/finger Fingers Motion Material Control Assessment design mode (g) (N) (kPa) (Ref.) AT Hand disabilities Grasping CPD Pneumatic (4) PAM 11 All except little Flexion-extension — 160 25 500 CLC Bending angle [57] AT Hand disabilities Grasping CPD Pneumatic (5) PAM 14 All Flexion-extension Elastomer 180 3 300 CLC Grasping force [127] 75 CLC RT Stroke survivors Grasping CPD Pneumatic (5) PAM 14 All Flexion-extension Latex —— — Bending angle [68] proportional 76 Thumb, index, Flexion-extension, RT Stroke survivors Grasping OPD Tendon-driven (5) PAM 9 Silicone KE-1300 T 120 12 — CLC Bending angle [72] and middle opposition/reposition 77 All except AT Stroke survivors Grasping CPD Tendon-driven (5) PAM 12 Flexion-extension Lycra — 16-17 — CLC Fingertip force ROM [49] thumb RT Stroke survivors Grasping CPD Tendon-driven (5) AAM-PAM 14 All Flexion-extension Elastomer >1000 —— CLC (PD) Fingertip force ROM [100] Flexion-extension, 79 Pneumatic Bending angle and AT-RT Hand disabilities Grasping CPD PAM 16 All opponent, adduction- Polyester 76 0.8 150 CLC [69] FREA (5) force output abduction 80 RTV-4234T4, RT Stroke survivors Grasping OPD Pneumatic (5) PAM 14 All Flexion-extension —— 50 CLC (PD) Bending angle [88] silicon RT Stroke survivors Grasping CPD Tendon-driven (5) PAM 14 All Flexion-extension Fabrics —— — — Virtual reality, FMA [40] AT Hand disabilities Grasping Semiopen Motor-tendon (5) PAM 14 All Flexion-extension Cotton fabric 600 —— On-off control Grasping force output [10] 83 Flexion-extension, AT Hand disabilities Grasping CPD Tendon-driven (5) PAM 12 All except little — 220 83 — CLC Grasping force ROM [90] opponent 84 Tendon-driven AT Heavy tasks Manipulation tasks CPD PAM 14 All Flexion-extension Rubber — 70 — CLC (PID) Force output [26] SMA (5) AT Hand disabilities Grasping, releasing CPD Pneumatic (5) PAM 14 All Flexion-extension Fabric 160 88.29 180 — EMG signals [104] 86 ROM, force output, RT Hand disabilities Grasping CPD Steel spring AAM-PAM 14 All Flexion-extension — 401 30.87 — CLC [91] EEG signals 87 EMG signals, grasping RT Hand disabilities Grasping CPD Pneumatic (5) PAM 14 All Flexion-extension — 150 40 300 OLC [58] forces 88 EMG signals, grasping AT Stroke survivors Grasping CPD Tendon-driven (5) AAM-PAM 10 All Flexion-extension Nylon 258 —— OLC [96] and lifting forces 89 Tendon-driven Teleoperation, RT Stroke survivors Grasping CPD PAM 14 All Flexion-extension —— — — — [103] SMA (5) time output 90 RTV-4234T4, Finger trajectories RT Stroke survivors Grasping CPD Pneumatic (5) PAM 14 All Flexion-extension —— 105 CLC (PD) [101] silicon and angle 91 Polyester and Griping force and AT Hand disabilities Grasping CPD Tendon-driven (5) PAM 14 All Flexion-extension —— — CLC [20] neoprene FMG signals Applied Bionics and Biomechanics 13 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 Year Figure 3: SEG developments in the last decade. adults. Thus, adjustable devices are recommended to have therapist availability, to the fact that clinical facilities are struggling to provide rehabilitation training, and to the the possibility to initiate an early SEG-based rehabilitation expensive costs of these health services. program since this is a common advice given by therapists, There are still significant challenges to face in soft exoglove no matter the dimensions of the patient’s hand. So far, SEG design. For instance, power supply approaches are still limited systems are able to accomplish full open-close fist, grasping, and tendon-driven actuation necessitates motors without lifting, and object release. Therefore, the systems reported heating problems, whereas hysteresis issues should be solved in literature encompass from 8 to 14 DOF. Moreover, SEG in pneumatic systems to increase actuation cycles and durabil- characterization could be developed to obtain more DOF in ity along with lightweight and portable power supplies. order to expand the workspace if needed. Regarding rehabilitation approaches, SEG systems must When soft exogloves are used, patient safety is a priority. be endowed to exert intensive and repetitive routines without Thus, human-machine interfaces with emergency buttons muscle fatigue and with minimal therapist assistance to excel and haptic feedback must be considered for harmless interac- above other rehabilitation options. SEG are a supportive aid tions [35, 128] as stated in Section 2 of this paper, and several safety strategies must be incorporated in every SEG system. that contributes to accelerated hand recovery by therapy pro- tocols. Nevertheless, to achieve a desired rehabilitation task, Moreover, SEG systems should not obstruct natural hand an active contribution from the patient is required to regain mobility and do not affect active ROM. Additionally, new strength, mobility, and ROM. Since the progress of each developments are expected to provide patients and thera- patient is variable, an AAM with time-triggered control could pists with useful information in order to evaluate patient progress. Furthermore, the capability to automatically adjust be implemented to regulate the input force of patients during rehabilitation processes, depending on their physical condi- the operation parameters as a function of the patient recov- tion. SEG systems must encourage patient participation but ery level is desirable. do not execute all the rehabilitation work. SEG self-manufacturing designs must ensure functional Several works have demonstrated that soft exogloves operation for home rehabilitation to provide low-cost sys- tems. These considerations could allow to improve SEG fea- have the potential to offer safe human-robot rehabilitation or assistance. However, new trends show that these two tasks tures as hours of operation, power consumption, cleaning, should be integrated into a unified system as it is reported by and maintenance. Since Bluetooth communications have [46, 92]. To accomplish integral rehabilitation, SEG designers been considered between SEG systems and control interfaces must consider that modular devices are expected to help [74], this or other communication systems must be part of new SEG devices when dealing with CLC strategies and for therapists and patients depending on the impairment or on the rehabilitation protocol. This will be satisfied by connect- rehabilitation or assistance data analysis. ing a soft exoglove device to a soft exosuit with a reliable and From this review, it can be pointed out that in recent robust platform (see, for instance, [28]). years, the development of SEG has grown significantly in SEG shortcomings were identified concerning different rehabilitation clinics and research groups. However, there is no comparison between research prototypes and those that hand sizes since most available systems are oriented towards Number of soft exogloves 14 Applied Bionics and Biomechanics criteria have been identified, classified, and established into have been already commercialized because the level of their technological maturity is different for each of them. Com- 2 function, 6 operation, and 5 usability criteria. mercialized SEG systems must have evolved from research This paper also provides 15 guidelines for SEG design, a prototypes. The main difference between these two types of detailed description of 91 SEG that have been analysed based devices is the one related to their technological maturity. on the aforementioned criteria, and a discussion that con- For instance, research prototypes can reach, in favorable siders different aspects in order to enhance future SEG cases, a technology readiness level (TRL) of 4 or 5 while com- developments. mercialized products have the highest TRL of 9 in China From this review, it is highlighted that patient safety [129, 130]. The evolution of a research prototype going from should be a priority characteristic during SEG operation, a 5 TRL to a certified product with 8 TRL and to a commer- and then, it should be guaranteed in every new SEG devel- cial product with a 9 TRL can take several years and require opment. This goal can be achieved by working closely with significant quantities of money. Moreover, medical devices a therapist, as recommended in [28], as well as incorporat- having official approvals or certifications as that of the Food ing safety in mechanical and electronical parts and in the and Drug Administration (FDA) or the Conformité Europé- programming of the SEG device. Moreover, safety stan- enne (CE) can be commercialized since they satisfy specific dards have been referenced to be considered in every requirements and standards while research prototypes focus, SEG development. mainly, on satisfying functional aspects. Then, it can be It has been remarked that several efforts have been made stated that commercialized medical devices are reliable due in terms of SEG designs. However, there is still room to to the fact that they have completed the product design cycle improve these devices. Then, this paper provides suggestions reaching the product life-cycle management, while research on patient safety, functional and continuous operation, prototypes have not begun the product development cycle friendly interaction, feedback information, and materials. or their industrial manufacture yet. Other areas to be explored include hybrid SEG systems New-generation products should seek for an affordable where new assembly techniques ensure force transmission trade-off between cost and benefit and include the possibility or the use of electroencephalography signals to monitor to perform assistance or rehabilitation therapy at home or in brain activity when SEG rehabilitation is performed. SEG specialized clinics to ensure that rehabilitation protocols, systems should be able to combine passive and active assis- defined by therapists, are efficiently executed. tance modes along with bilateral training to enhance recov- SEG designs should provide acceptable appearance, com- ery processes and to encourage patients. The mentioned fort, and functionality to patients. Hence, it is highly recom- SEG design criteria provide perfectible guidelines to improve mended that SEG systems consider accessible technologies their performance and represent a basis to develop SEG that could, additionally, create dynamic environments where robust designs. patients can have pleasant therapy sessions. SEG require materials with appearance and elastic modulus similar to Abbreviations human tissues. Thus, smart polymers represent the primary current choice due to their biomimetic qualities to develop AAM: Active assistance mode lightweight devices with modular OPD [128]. Besides, elasto- ADL: Activities of daily living mers have been shown to be compliant wearable components ARM: Active resistive mode with the ability to vary their form and increase the ROM ASI: Ashworth spasticity index based on the shape of the human hand. AT: Assistance tasks Modularity plays a significant role when dealing with BBT: Box and block test maintenance aspects of SEG systems as well as with costs CLC: Closed-loop control and should be considered in new SEG developments. Besides, CE: Conformité Européenne modularity can play a significant role when dealing with CMC: Carpometacarpal rehabilitation of different fingers or DOF. Regarding porta- CP: Cerebral palsy bility in new SEG developments, minimizing the dependence CPD: Closed palm design of energy sources becomes a challenge that must be DOF: Degrees of freedom addressed by researchers and engineers. EMG: Electromyography It has become clear that a SEG device that allows adapta- FDA: Food and Drug Administration tion (customization) to a larger number of patients without FEM: Finite element method the need for component replacements will be preferable to FIM: Functional independence measure another system that only works for a certain size of hands. FMA: Fugl-Meyer assessment FMG: Force myography FREA: Fiber reinforced elastomer actuators 5. Conclusions GUI: Graphical user interface IP: Interphalangeal JTT: Jebsen-Taylor hand test Scientific and technical communications concerning wear- able SEG for hand rehabilitation and assistance tasks applied MAS: Modified Ashworth scale to stroke survivors or people with hand disabilities have been MCP: Metacarpophalangeal extensively reviewed and reported in this paper. SEG design MIT: Motricity index test Applied Bionics and Biomechanics 15 [9] Z. Yue, X. Zhang, and J. Wang, “Hand Rehabilitation Robot- NHPT: Nine-hole peg test ics on Poststroke Motor Recovery,” Behavioural Neurology, OLC: Open-loop control vol. 2017, Article ID 3908135, 20 pages, 2017. OPD: Open palm design [10] R. Ismail, M. Ariyanto, T. Hidayat, and J. D. Setiawan, PAM: Passive assistance mode “Design of fabric-based soft robotic glove for hand function PD: Proportional-derivative assistance,” in 2019 6th International Conference on Informa- PID: Proportional-integral-derivative tion Technology, Computer and Electrical Engineering (ICI- PT: Purdue pegboard test TACEE), pp. 1–5, Semarang, Indonesia, Indonesia, 2019. ROM: Range of motion [11] P. Polygerinos, K. C. Galloway, E. Savage, M. Herman, RT: Robot rehabilitation K. O’Donnell, and C. J. Walsh, “Soft robotic glove for hand SCI: Spinal cord injury rehabilitation and task specific training,” in 2015 IEEE Inter- SEG: Soft exogloves national Conference on Robotics and Automation (ICRA), SMA: Shape memory alloys pp. 2913–2919, Seattle, WA, USA, 2015. RTV: Room-temperature-vulcanizing [12] D. Rus and M. T. Tolley, “Design, fabrication and control of TPU: Thermoplastic polyurethane soft robots,” Nature, vol. 521, no. 7553, pp. 467–475, 2015. TRL: Technology readiness level. [13] T. Noritsugu, M. Takaiwa, and D. Sasaki, “Power assist wear driven with pneumatic rubber artificial muscles,” in 2008 Conflicts of Interest 15th International Conference on Mechatronics and Machine Vision in Practice, pp. 539–544, Auckland, New Zealand, The authors declare that there is no conflict of interest regarding the publication of this paper. [14] H. In, K. J. Cho, K. Kim, and B. Lee, “Jointless structure and under-actuation mechanism for compact hand exoskeleton,” Acknowledgments in 2011 IEEE International Conference on Rehabilitation Robotics, pp. 1–6, Zurich, Switzerland, 2011. The authors thank CONACYT and Universidad Autón- [15] L. Connelly, Y. Jia, M. L. Toro, M. E. Stoykov, R. V. Kenyon, oma del Estado de México for all their support. Funding and D. G. Kamper, “A Pneumatic Glove and Immersive Vir- supporting this research was provided by CONACYT tual Reality Environment for Hand Rehabilitative Training (406476) and Universidad Autónoma del Estado de México After Stroke,” IEEE Transactions on Neural Systems and (5015-2020CIA). Rehabilitation Engineering, vol. 18, no. 5, pp. 551–559, 2010. [16] Y. Kadowaki, T. Noritsugu, M. Takaiwa, D. Sasaki, and References M. Kato, “Development of Soft Power-Assist Glove and Con- trol Based on Human Intent,” Journal of Robotics and Mecha- [1] H. K. Yap, J. H. Lim, J. C. H. Goh, and C.-H. Yeow, “Design of tronics, vol. 23, no. 2, pp. 281–291, 2011. a Soft Robotic Glove for Hand Rehabilitation of Stroke [17] U. Jeong, H.-K. In, and K.-J. Cho, “Implementation of various Patients With Clenched Fist Deformity Using Inflatable Plas- control algorithms for hand rehabilitation exercise using tic Actuators,” Journal of Medical Devices, vol. 10, no. 4, 2016. wearable robotic hand,” Intelligent Service Robotics, vol. 6, [2] V. L. Feigin, B. Norrving, and G. A. Mensah, “Global burden no. 4, pp. 181–189, 2013. of stroke,” Circulation Research, vol. 120, no. 3, pp. 439–448, 2017. [18] H. K. Yap, J. H. Lim, F. Nasrallah, J. C. H. Goh, and R. C. H. Yeow, “A soft exoskeleton for hand assistive and reha- [3] A. G. Thrift, T. Thayabaranathan, G. Howard et al., “Global bilitation application using pneumatic actuators with vari- stroke statistics,” International Journal of Stroke, vol. 12, able stiffness,” in 2015 IEEE International Conference on no. 1, pp. 13–32, 2017. Robotics and Automation (ICRA), pp. 4967–4972, Seattle, [4] B. Cherian, C. Dominic, G. Vysakh, and K. R. Vishakh, “Exo- WA, USA, 2015. Glove: A Soft Wearable Robotic Hand for Stroke Survivors,” [19] H. In, B. B. Kang, M. Sin, and K. J. Cho, “Exo-Glove: A Wear- International Research Journal of Engineering and Technol- able Robot for the Hand with a Soft Tendon Routing System,” ogy, vol. 5, no. 5, pp. 3358–3364, 2018. IEEE Robotics and Automation Magazine, vol. 22, no. 1, [5] B. Radder, G. B. Prange-Lasonder, A. I. R. Kottink et al., pp. 97–105, 2015. “Home rehabilitation supported by a wearable soft-robotic [20] J. Fajardo, A. R. Neto, W. Silva, M. Gomes, E. Fujiwara, and device for improving hand function in older adults: A pilot randomized controlled trial,” PLoS One, vol. 14, no. 8, article E. Rohmer, “A wearable robotic glove based on optical FMG driven controller,” in 2019 IEEE 4th International Con- e0220544, 2019. ference on Advanced Robotics and Mechatronics (ICARM), [6] B. Radder, “General Introduction,” The Wearable Hand pp. 81–86, Toyonaka, Japan, Japan, 2019. Robot, pp. 10–21, 2018. [21] B. B. Kang, H. Lee, H. In, U. Jeong, J. Chung, and K. J. Cho, [7] A. Stilli, A. Cremoni, M. Bianchi et al., “AirExGlove — A “Development of a polymer-based tendon-driven wearable novel pneumatic exoskeleton glove for adaptive hand rehabil- robotic hand,” in 2016 IEEE International Conference on itation in post-stroke patients,” in IEEE International Confer- Robotics and Automation (ICRA), pp. 3750–3755, Stockholm, ence on Soft Robotics (RoboSoft), vol. 2018, pp. 579–584, Sweden, 2016. Livorno, Italy, 2018. [8] P. Polygerinos, S. Lyne, Z. Wang et al., “Towards a soft pneu- [22] B. B. Kang, H. Choi, H. Lee, and K. J. Cho, “Exo-Glove Poly matic glove for hand rehabilitation,” in 2013 IEEE/RSJ Inter- II: A Polymer-Based Soft Wearable Robot for the Hand with national Conference on Intelligent Robots and Systems, a Tendon-Driven Actuation System,” Soft Robotics, vol. 6, pp. 1512–1517, Tokyo, Japan, 2013. no. 2, pp. 214–227, 2019. 16 Applied Bionics and Biomechanics [38] P. Bernocchi, C. Mulè, F. Vanoglio, G. Taveggia, A. Luisa, and [23] S. S. Yun, B. B. Kang, and K. J. Cho, “Exo-Glove PM: An Eas- ily Customizable Modularized Pneumatic Assistive Glove,” S. Scalvini, “Home-based hand rehabilitation with a robotic IEEE Robotics and Automation Letters, vol. 2, no. 3, glove in hemiplegic patients after stroke: a pilot feasibility pp. 1725–1732, 2017. study,” Topics in Stroke Rehabilitation, vol. 25, pp. 114–119, [24] H. Kazerooni, D. Fairbanks, A. Chen, and G. Shin, “The magic glove,” in 2004 in IEEE International Conference on [39] J.-H. Shin, M.-Y. Kim, J.-Y. Lee et al., “Effects of virtual reality-based rehabilitation on distal upper extremity func- Robotics and Automation, 2004. Proceedings. ICRA '04. 2004, vol. 1, pp. 757–763, New Orleans, LA, USA, USA, 2004. tion and health-related quality of life: a single-blinded, ran- domized controlled trial,” Journal of NeuroEngineering and [25] M. Diftler, C. A. Ihrke, L. B. Bridgwater et al., “RoboGlove - A Rehabilitation, vol. 13, no. 1, 2016. Robonaut Derived Multipurpose Assistive Device,” in Inter- [40] B. Radder, G. B. Prange-Lasonder, A. I. R. Kottink, national Conference on Robotics and Automation, Hong A. Melendez-Calderon, J. H. Buurke, and J. S. Rietman, “Fea- Kong, China, 2014. sibility of a wearable soft-robotic glove to support impaired [26] Á. Villoslada, C. Rivera, N. Escudero, F. Martín, D. Blanco, hand function in stroke patients,” Journal of Rehabilitation and L. Moreno, “Hand Exo-Muscular System for Assisting Medicine, vol. 50, no. 7, pp. 598–606, 2018. Astronauts During Extravehicular Activities,” Soft Robotics, [41] H. K. Yap, J. H. Lim, F. Nasrallah, F. Z. Low, J. C. H. Goh, and vol. 6, no. 1, pp. 21–37, 2019. R. C. H. Yeow, “MRC-glove: A fMRI compatible soft robotic [27] P. Ohta, L. Valle, J. King et al., “Design of a Lightweight glove for hand rehabilitation application,” in 2015 IEEE Inter- Soft Robotic Arm Using Pneumatic Artificial Muscles and national Conference on Rehabilitation Robotics (ICORR), Inflatable Sleeves,” Soft Robotics, vol. 5, no. 2, pp. 204– pp. 735–740, Singapore, Singapore, 2015. 215, 2018. [42] Y. Jiang, D. Chen, J. Que, Z. Liu, Z. Wang, and Y. Xu, “Soft [28] J.-M. Dávila-Vilchis, J. C. Á. V. LAZ-Avilés, and A. H. robotic glove for hand rehabilitation based on a novel fabrica- Vilchis-González, “Design Methodology for Soft Wearable tion method,” in 2017 IEEE International Conference on Devices—The MOSAR Case,” Applied Sciences, vol. 9, no. 22, Robotics and Biomimetics (ROBIO), pp. 817–822, Macau, p. 4727, 2019. China, 2017. [29] M. Troncossi, M. Mozaffari-Foumashi, and V. Parenti-Cas- [43] H. K. Yap, B. W. K. Ang, J. H. Lim, J. C. H. Goh, and C. H. telli, “An original classification of rehabilitation hand exo- Yeow, “A fabric-regulated soft robotic glove with user intent skeletons,” Journal of Robotics and Mechanical Engineering detection using EMG and RFID for hand assistive application,” Research, vol. 1, no. 4, pp. 17–29, 2016. in 2016 IEEE International Conference on Robotics and Auto- [30] Y. Chen, S. Le, Q. C. Tan, O. Lau, F. Wan, and C. Song, “A mation (ICRA),pp.3537–3542, Stockholm, Sweden, 2016. lobster-inspired robotic glove for hand rehabilitation,” in [44] H. K. Yap, J. H. Lim, F. Nasrallah, and C.-H. Yeow, “Design 2017 IEEE International Conference on Robotics and Automa- and Preliminary Feasibility Study of a Soft Robotic Glove tion (ICRA), pp. 4782–4787, Singapore, Singapore, 2017. for Hand Function Assistance in Stroke Survivors,” Frontiers [31] S. Biggar and W. Yao, “Design and Evaluation of a Soft and in Neuroscience, vol. 11, 2017. Wearable Robotic Glove for Hand Rehabilitation,” IEEE [45] F. Vanoglio, A. Luisa, F. Garofali, and C. Mora, “Evaluation Transactions on Neural Systems and Rehabilitation Engineer- of the effectiviness of Glohera (hand rehabilitation glove) ing, vol. 24, no. 10, pp. 1071–1080, 2016. on hemiplegic patients. Pilot study,” in XIII Congress of Ital- [32] J. Brackenridge, L. V. Bradnam, S. Lennon, J. J. Costi, and ian Society of Neurorehabilitation, pp. 18–20, Bari, Italy, D. A. Hobbs, “A Review of Rehabilitation Devices to Promote Upper Limb Function Following Stroke,” Neuroscience and [46] G. B. Prange-Lasonder, B. Radder, A. I. R. Kottink, Biomedical Engineering, vol. 4, no. 1, pp. 25–42, 2016. A. Melendez-Calderon, J. H. Buurke, and J. S. Rietman, [33] G. A. Prieto, A. I. Andrés, A. Venkatakrishnan, W. Q. Malik, “Applying a soft-robotic glove as assistive device and training V. Dietz, and W. Z. Rymer, “Rehabilitation Technologies for tool with games to support hand function after stroke: Pre- Spinal Injury,” in Emerging Therapies in Neurorehabilitation liminary results on feasibility and potential clinical impact,” II, pp. 65–85, Springer, 2016. in 2017 International Conference on Rehabilitation Robotics [34] K. Serbest, S. Ateş, and A. H. A. Stienen, “Design of an exer- (ICORR), pp. 1401–1406, London, UK, 2017. cise glove for hand rehabilitation using spring mechanism,” [47] H. Kim, A. Lee, Y. I. Shin et al., “Effects of digital smart glove in 2016 20th National Biomedical Engineering Meeting system on motor recovery of upper extremity in subacute (BIYOMUT), pp. 1–5, Izmir, Turkey, 2016. stroke patients,” Annals of Physical and Rehabilitation Medi- cine, vol. 61, p. e28, 2018. [35] H. Cao and D. Zhang, “Soft robotic glove with integrated sEMG sensing for disabled people with hand paralysis,” in [48] M. Haghshenas-Jaryani, W. Carrigan, C. Nothnagle, and M. B. 2016 IEEE International Conference on Robotics and Biomi- J. Wijesundara, “Sensorized soft robotic glove for continuous metics (ROBIO), pp. 714–718, Qingdao, China, 2016. passive motion therapy,” in 2016 6th IEEE International Con- ference on Biomedical Robotics and Biomechatronics (BioRob), [36] A. Rodigari, A. Zancan, I. Jedrychowska, and I. Springhetti, pp. 815–820, Singapore, Singapore, 2016. “Effectiveness of the treatment of hemiplegic patient’s hand with robotic rehabilitation glove “gloreha”: preliminary [49] S. Toochinda and W. Wannasuphoprasit, “Design and results,” in The 42nd National Congress Of The Italian Society Development of an Assistive Hand Device for Enhancing Of Physical And Rehabilitative Medicine, pp. 142–145, Compatibility and Comfortability,” in 2018 2nd International Torino, Italy, 2014, SIMFER. Conference on Engineering Innovation (ICEI), pp. 1–6, Bang- kok, Thailand, 2018. [37] P. Maeder-York, T. Clites, E. Boggs et al., “Biologically Inspired Soft Robot for Thumb Rehabilitation1,” Journal of [50] P. Polygerinos, K. C. Galloway, S. Sanan, M. Herman, and Medical Devices, vol. 8, no. 2, 2014. C. J. Walsh, “EMG controlled soft robotic glove for assistance Applied Bionics and Biomechanics 17 during activities of daily living,” in 2015 IEEE International [65] Y. Tsuji, “A safety review of medical device robots in Japan,” Conference on Rehabilitation Robotics (ICORR), pp. 55–60, sia Pacific Journal of Health Law &Ethics, vol. 12, no. 1, pp. 1– Singapore, Singapore, 2015. 27, 2018. [51] P. Polygerinos, Z. Wang, K. C. Galloway, R. J. Wood, and C. J. [66] P. Barattini, V. Federico, G. S. Virk, and T. Haidegger, Human–Robot Interaction, Taylor & F. Boca Raton: CRC Walsh, “Soft robotic glove for combined assistance and at- home rehabilitation,” Robotics and Autonomous Systems, Press, New York NY USA, 2019. vol. 73, pp. 135–143, 2015. [67] Z. Sun, Z. Guo, and W. Tang, “Design of wearable hand reha- bilitation glove with soft hoop-reinforced pneumatic actua- [52] M. Xiloyannis, L. Cappello, B. Khanh Dinh, C. W. Antuvan, tor,” Journal of Central South University, vol. 26, no. 1, and L. Masia, “Design and Preliminary Testing of a Soft Exo- pp. 106–119, 2019. suit for Assisting Elbow Movements and Hand Grasping,” in Converging Clinical and Engineering Research on Neuroreh- [68] H. Li and L. Cheng, “Preliminary study on the design and abilitation II, vol. 15, pp. 557–561, Springer, 2017. control of a pneumatically-actuated hand rehabilitation device,” in 2017 32nd Youth Academic Annual Conference [53] M. Xiloyannis, L. Cappello, D. B. Khanh, S.-C. Yen, and of Chinese Association of Automation (YAC), pp. 860–865, L. Masia, “Modelling and design of a synergy-based actuator Hefei, 2017. for a tendon-driven soft robotic glove,” in 2016 6th IEEE [69] T. Jiralerspong, K. H. L. Heung, R. K. Y. Tong, and Z. Li, “A International Conference on Biomedical Robotics and Biome- chatronics (BioRob), pp. 1213–1219, Singapore, Singapore, novel soft robotic glove for daily life assistance,” in 2018 7th IEEE International Conference on Biomedical Robotics and Biomechatronics (Biorob), pp. 671–676, Enschede, 2018. [54] C. Jiménez, R. Mora, J. Pérez, and K. Quirós, “Nitiglove: [70] L. Wang, S. G. Nurzaman, and F. Iida, “Soft-material robot- Nitinol-Driven Robotic Glove Used to Assist Therapy for ics,” Foundations and Trends in Robotics, vol. 5, no. 3, Hand Mobility Recovery,” in Wearable Robotics: Challenges pp. 191–259, 2017. and Trends, pp. 379–383, Springer, 2017. [71] B. Wang, A. McDaid, M. Biglari-Abhari, and K. C. Aw, [55] A. Hadi, K. Alipour, S. Kazeminasab, and M. Elahinia, “ASR “Design and development of a glove for post-stroke hand glove: A wearable glove for hand assistance and rehabilitation rehabilitation,” in 2017 IEEE International Conference on using shape memory alloys,” Journal of Intelligent Material Advanced Intelligent Mechatronics (AIM), pp. 1047–1051, Systems and Structures, vol. 29, no. 8, pp. 1575–1585, 2017. Munich, 2017. [56] S. Sridar, P. H. Nguyen, M. Zhu, Q. P. Lam, and [72] D. H. Kim and H. S. Park, “Cable actuated dexterous P. Polygerinos, “Development of a soft-inflatable exosuit for (CADEX) glove for effective rehabilitation of the hand for knee rehabilitation,” in 2017 IEEE/RSJ International Confer- patients with neurological diseases,” in 2018 IEEE/RSJ Inter- ence on Intelligent Robots and Systems (IROS), pp. 3722– national Conference on Intelligent Robots and Systems (IROS), 3727, Vancouver, BC, Canada, 2017. Madrid, Spain, 2018. [57] H. Al-Fahaam, S. Davis, S. Nefti-Meziani, and T. Theodoridis, [73] M. A. Delph, S. A. Fischer, P. W. Gauthier, C. H. M. Luna, “Novel soft bending actuator-based power augmentation hand E. A. Clancy, and G. S. Fischer, “A soft robotic exomuscula- exoskeleton controlled by human intention,” Intelligent Service ture glove with integrated sEMG sensing for hand rehabilita- Robotics,vol. 11,no.3, pp.247–268, 2018. tion,” in 2013 IEEE 13th International Conference on [58] H. Al-Fahaam, S. Davis, and S. Nefti-Meziani, “The design Rehabilitation Robotics (ICORR), pp. 1–7, Seattle, WA, 2013. and mathematical modelling of novel extensor bending [74] A. Mohammadi, J. Lavranos, P. Choong, and D. Oetomo, pneumatic artificial muscles (EBPAMs) for soft exoskele- “Flexo-glove: a 3D printed soft exoskeleton robotic glove for tons,” Robotics and Autonomous Systems, vol. 99, pp. 63–74, impaired hand rehabilitation and assistance,” in 2018 40th Annual International Conference of the IEEE Engineering [59] M. Xiloyannis, L. Cappello, K. D. Binh, C. W. Antuvan, and in Medicine and Biology Society (EMBC), pp. 2120–2123, L. Masia, “Preliminary design and control of a soft exosuit Honolulu, HI, 2018. for assisting elbow movements and hand grasping in activi- [75] H. K. Yap, P. M. Khin, T. H. Koh et al., “A fully fabric-based ties of daily living,” Journal of Rehabilitation and Assistive, bidirectional soft robotic glove for assistance and rehabilita- vol. 4, article 205566831668031, 2017. tion of hand impaired patients,” IEEE Robotics and Automa- [60] H. Abidi and M. Cianchetti, “On intrinsic safety of soft tion Letters, vol. 2, no. 3, pp. 1383–1390, 2017. robots,” Frontiers in Robotics and AI, vol. 4, no. 5, 2017. [76] H. K. Yap, J. C. H. Goh, and R. C. H. Yeow, “Design and char- [61] S. Jadhav, V. Kannanda, B. Kang, M. T. Tolley, and J. P. acterization of soft actuator for hand rehabilitation applica- Schulze, “Soft robotic glove for kinesthetic haptic feedback tion,” in 6th European Conference of the International in virtual reality environments,” Electronic Imaging, Federation for Medical and Biological Engineering, Springer, vol. 2017, no. 3, pp. 19–24, 2017. pp. 367–370, Dubrovnik, 2015. [62] A. Avižienis, J. C. Laprie, B. Randell, and C. Landwehr, “Basic [77] H. K. Yap, J. H. Lim, F. Nasrallah, J. C. H. Goh, and C.- concepts and taxonomy of dependable and secure comput- H. Yeow, “Characterisation and evaluation of soft elastomeric ing,” IEEE Transactions on Dependable and Secure Comput- actuators for hand assistive and rehabilitation applications,” ing, vol. 1, no. 1, pp. 11–33, 2004. Journal of Medical Engineering & Technology, vol. 40, no. 4, [63] J. Guiochet, M. Machin, and H. Waeselynck, “Safety-critical pp. 199–209, 2016. advanced robots: a survey,” Robotics and Autonomous Sys- [78] H. K. Yap, F. Sebastian, C. Wiedeman, and C. H. Yeow, tems, vol. 94, pp. 43–52, 2017. “Design and characterization of low-cost fabric-based flat [64] K. Chinzei, “Safety of surgical robots and IEC 80601-2-77: the pneumatic actuators for soft assistive glove application,” in first international standard for surgical robots,” Acta Poly- 2017 International Conference on Rehabilitation Robotics technica Hungarica, vol. 16, no. 8, pp. 171–184, 2019. (ICORR), pp. 1465–1470, London, 2017. 18 Applied Bionics and Biomechanics ference on Rehabilitation Robotics (ICORR), pp. 502–507, Sin- [79] A. F. Hassanin, D. Steve, and N. M. Samia, “A novel, soft, bending actuator for use in power assist and rehabilitation gapore, 2015. exoskeletons,” in 2017 IEEE/RSJ International Conference [93] C. D. Takahashi, L. Der-Yeghiaian, V. Le, R. R. Motiwala, and on Intelligent Robots and Systems (IROS), pp. 533–538, Van- S. C. Cramer, “Robot-based hand motor therapy after stroke,” couver, BC, 2017. Brain, vol. 131, no. 2, pp. 425–437, 2008. [80] P. Polygerinos, Z. Wang, J. T. B. Overvelde et al., “Modeling [94] T. Proietti, V. Crocher, A. Roby-Brami, and N. Jarrassé, of soft fiber-reinforced bending actuators,” IEEE Transac- “Upper-limb robotic exoskeletons for neurorehabilitation: a tions on Robotics, vol. 31, no. 3, pp. 778–789, 2015. review on control strategies,” IEEE Reviews in Biomedical [81] B. Wang, K. C. Aw, M. Biglari-Abhari, and A. McDaid, Engineering, vol. 9, pp. 4–14, 2016. “Design and fabrication of a fiber-reinforced pneumatic [95] C. M. L. Hughes, P. Tommasino, A. Budhota, and bending actuator,” in 2016 IEEE International Conference D. Campolo, “Upper extremity proprioception in healthy on Advanced Intelligent Mechatronics (AIM), pp. 83–88, aging and stroke populations, and the effects of therapist- Banff, 2016. and robot-based rehabilitation therapies on proprioceptive function,” Frontiers in Human Neuroscience, vol. 9, 2015. [82] Y. Jiang, D. Chen, P. Liu et al., “Fishbone-inspired soft robotic glove for hand rehabilitation with multi-degrees-of-free- [96] M. K. Burns, D. Pei, and R. Vinjamuri, “Myoelectric control dom,” in 2018 IEEE International Conference on Soft Robotics of a soft hand exoskeleton using kinematic synergies,” (RoboSoft), pp. 394–399, Livorno, 2018. vol. 13, Tech. Rep. 6, IEEE Transactions on Biomedical Cir- cuits and Systems, 2019. [83] H. Lee, B. B. Kang, H. In, and K.-J. Cho, Design improvement of a polymer-based tendon-driven wearable robotic hand (exo- [97] H. K. Yap, A. Mao, J. C. H. Goh, and C. H. Yeow, “Design of a glove poly), vol. 16 of Wearable Robotics: Challenges and wearable FMG sensing system for user intent detection dur- Trends, , no. 1, 2017Springer, 2017. ing hand rehabilitation with a soft robotic glove,” in 2016 6th IEEE International Conference on Biomedical Robotics [84] M. Nilsson, J. Ingvast, J. Wikander, and H. von Holst, “The and Biomechatronics (BioRob), pp. 781–786, Singapore, 2016. soft extra muscle system for improving the grasping capabil- ity in neurological rehabilitation,” in 2012 IEEE-EMBS Con- [98] N. F. M. Nawi, A. H. K. Ching, N. Ibrahim et al., “Design and ference on Biomedical Engineering and Sciences, pp. 412– development approach of smart glove for post stroke rehabil- 417, Langkawi, 2012. itation,” International Journal of Engineering & Technology, vol. 7, no. 4, pp. 78–82, 2018. [85] F. L. Hammond, Y. Mengüç, and R. J. Wood, “Toward a modular soft sensor-embedded glove for human hand [99] S. Sasidharan, Smart glove: an assistive device to enhanced motion and tactile pressure measurement,” in 2014 IEEE/RSJ recovery of hand function during motor rehabilitation, [M.S. International Conference on Intelligent Robots and Systems, thesis], Arizona State University, 2015. pp. 4000–4007, Chicago, IL, 2014. [100] G. Rudd, L. Daly, V. Jovanovic, and F. Cuckov, “A low-cost [86] F. Schmitt, O. Piccin, L. Barbé, and B. Bayle, “Soft robots soft robotic hand exoskeleton for use in therapy of limited manufacturing: a review,” Frontiers in Robotics and AI, hand-motor function,” vol. 9, Tech. Rep. 18, Applied Sci- vol. 5, pp. 84–101, 2018. ences, 2019. [101] M. Haghshenas-Jaryani, C. Pande, and B. J. Muthu Wijesun- [87] P. Polygerinos, N. Correll, S. A. Morin et al., “Soft Robotics: Review of Fluid‐Driven Intrinsically Soft Devices; dara, “Soft robotic bilateral hand rehabilitation system for fine motor learning,” 2019 IEEE 16th Intertnational Confer- Manufacturing, Sensing, Control, and Applications in Human‐Robot Interaction,” Advanced Engineering Mate- ence on Rehabilitation Robotics (ICORR), pp. 337–342, Toronto, ON, 2019. rials, vol. 19, no. 12, pp. 1–22, 2017. [102] T. George Thuruthel, Y. Ansari, E. Falotico, and C. Laschi, [88] M. Haghshenas-Jaryani, C. Nothnagle, R. M. Patterson, “Control strategies for soft robotic manipulators: a survey,” N. Bugnariu, and M. B. J. Wijesundara, “Soft robotic rehabil- Soft Robotics, vol. 5, no. 2, pp. 149–163, 2018. itation exoskeleton (rehab glove) for hand therapy,” in Vol- ume 3: 19th International Conference on Advanced Vehicle [103] D. Popescu, M. Ivanescu, R. Popescu, A. Petrisor, L. C. Technologies; 14th International Conference on Design Educa- Popescu, and A. M. Bumbea, “Post-stroke hand rehabilitation tion; 10th Frontiers in Biomedical Devices, Cleveland, OH, using a wearable robotic glove,” in Smart Innovation, Systems and Technologies, pp. 259–268, Tenerife, Spain, 2016. [89] B. W. K. Ang and C. H. Yeow, “Print-it-Yourself (PIY) glove: [104] J. Nassour and F. Hamker, “Enfolded textile actuator for soft a fully 3D printed soft robotic hand rehabilitative and assis- wearable robots,” in 2019 IEEE International Conference on tive exoskeleton for stroke patients,” in 2017 IEEE/RSJ Inter- Cyborg and Bionic Systems (CBS), pp. 60–65, Munich, Ger- national Conference on Intelligent Robots and Systems (IROS), many, 2019. pp. 1219–1223, Vancouver, 2017. [105] F. Vanoglio, P. Bernocchi, C. Mulè et al., “Feasibility and effi- [90] C. G. Rose and M. K. O'Malley, “Hybrid rigid-soft hand exo- cacy of a robotic device for hand rehabilitation in hemiplegic skeleton to assist functional dexterity,” IEEE Robotics and stroke patients: a randomized pilot controlled study,” Clinical Automation Letters, vol. 4, no. 1, pp. 73–80, 2019. Rehabilitation, vol. 31, no. 3, pp. 351–360, 2017. [91] M. Li, B. He, Z. Liang et al., “An attention-controlled hand [106] H. K. Yap, N. Kamaldin, J. H. Lim, F. A. Nasrallah, J. C. H. exoskeleton for the rehabilitation of finger extension and flex- Goh, and C. H. Yeow, “A magnetic resonance compatible soft ion using a rigid-soft combined mechanism,” Frontiers in wearable robotic glove for hand rehabilitation and brain Neurorobotics, vol. 13, no. 34, pp. 1–13, 2019. imaging,” IEEE Transactions on Neural Systems and Rehabil- itation Engineering, vol. 25, no. 6, pp. 782–793, 2017. [92] B. Radder, A. I. R. Kottink, N. van der Vaart et al., “User- centred input for a wearable soft-robotic glove supporting [107] D. Popov, I. Gaponov, and J. H. Ryu, “Portable exoskeleton hand function in daily life,” in 2015 IEEE International Con- glove with soft structure for hand assistance in activities of Applied Bionics and Biomechanics 19 Communication and Electronics Systems (ICCES), pp. 341– daily living,” IEEE/ASME Transactions on Mechatronics, vol. 22, no. 2, pp. 865–875, 2017. 345, Coimbatore, India, 2017. [108] K. Shiota, S. Kokubu, T. V. J. Tarvainen et al., “Enhanced [123] K. O. Thielbar, K. M. Triandafilou, H. C. Fischer et al., Kapandji test evaluation of a soft robotic thumb rehabilitation “Benefits of using a voice and EMG-driven actuated glove device by developing a fiber-reinforced elastomer-actuator to support occupational therapy for stroke survivors,” IEEE Transactions on Neural Systems and Rehabilitation Engi- based 5-digit assist system,” Robotics and Autonomous Systems, vol. 111, pp. 20–30, 2019. neering, vol. 25, no. 3, pp. 297–305, 2017. [109] B. Kim, H. In, D. Y. Lee, and K. J. Cho, “Development and [124] K. H. L. Heung, R. K. Y. Tong, A. T. H. Lau, and Z. Li, “Robotic glove with soft-elastic composite actuators for assessment of a hand assist device: GRIPIT,” Journal of Neu- roengineering and Rehabilitation, vol. 14, no. 1, 2017. assisting activities of daily living,” Soft Robotics, vol. 6, no. 2, pp. 289–304, [110] A. L. Coffey, D. J. Leamy, and T. E. Ward, “A novel BCI- controlled pneumatic glove system for home-based neuror- [125] L. Cappello, J. T. Meyer, K. C. Galloway et al., “Assisting hand ehabilitation,” in 2014 36th Annual International Conference function after spinal cord injury with a fabric-based soft of the IEEE Engineering in Medicine and Biology Society, robotic glove,” Journal of NeuroEngineering and Rehabilita- pp. 3622–3625, Chicago, IL, 2014. tion, vol. 15, no. 1, 2018. [111] A. M. Gordon, Y. C. Hung, M. Brandao et al., “Bimanual [126] P.I.Daiya, “Daiya glove,” 2015, https://exoskeletonreport training and constraint-induced movement therapy in chil- .com/product/daiya-glove/. dren with hemiplegic cerebral palsy,” Neurorehabilitation [127] K. Toya, T. Miyagawa, and Y. Kubota, “Power-assist glove and Neural Repair, vol. 25, no. 8, pp. 692–702, 2011. operated by predicting the grasping mode,” Journal of System [112] L. Gerez, J. Chen, and M. Liarokapis, “On the development of Design and Dynamics, vol. 5, no. 1, pp. 94–108, 2011. adaptive, tendon-driven, wearable exo-gloves for grasping [128] C.-Y. Chu and R. M. Patterson, “Soft robotic devices for hand capabilities enhancement,” IEEE Robotics and Automation rehabilitation and assistance: a narrative review,” Journal of Letters, vol. 4, no. 2, pp. 422–429, 2019. NeuroEngineering and Rehabilitation, vol. 15, no. 1, pp. 9– [113] Glohera, “Hand rehabilitation glove,” 2019, http://www 14, 2018. .gloreha.com/?lang=it. [129] Z. Ouyang, P. Zhou, J. Du, and H. Chi, “Modified technology [114] D. Ryu, K.-W. Moon, H. Nam et al., “Micro hydraulic system readiness level for advanced medical devices in China,” West- using slim artificial muscles for a wearable haptic glove,” in ern Pacific Region Index Medicus, vol. 28, no. 3, pp. 217–221, 2008 IEEE/RSJ International Conference on Intelligent Robots 2015. and Systems, pp. 3028–3033, Nice, 2008. [130] ISO, “ISO 16290:2013,” 2013, https://www.iso.org/standard/ [115] Neofect, “Hand assistance glove,” 2019, http://www.neofect 56064.html. .com/en/product/rapael/. [116] J. B. Chossat, Y. Tao, V. Duchaine, and Y. L. Park, “Wearable soft artificial skin for hand motion detection with embedded microfluidic strain sensing,” in 2015 IEEE International Con- ference on Robotics and Automation (ICRA), pp. 2568–2573, Seattle, Washington, 2015. [117] B. Radder, G. B. Prange-Lasonder, A. I. R. Kottink et al., “A wearable soft-robotic glove enables hand support in ADL and rehabilitation: a feasibility study on the assistive func- tionality,” Journal of Rehabilitation and Assistive Technolo- gies Engineering, vol. 3, 2016. [118] B. Radder, G. B. Prange-Lasonder, A. I. R. Kottink et al., “The effect of a wearable soft-robotic glove on motor function and functional performance of older adults,” Assistive Technology, vol. 32, pp. 1–27, 2020. [119] J. Yi, Z. Shen, C. Song, and Z. Wang, “A soft robotic glove for hand motion assistance,” in IEEE International Conference on Real-time Computing and Robotics (RCAR), pp. 111–116, Angkor, 2016. [120] J. Yi, X. Chen, and Z. Wang, “A three-dimensional-printed soft robotic glove with enhanced ergonomics and force capa- bility,” IEEE Robotics and Automation Letters, vol. 3, no. 1, pp. 242–248, 2018. [121] Z. Yao, C. Linnenberg, A. Argubi-Wollesen, R. Weidner, and J. P. Wulfsberg, “Biomimetic design of an ultra-compact and light-weight soft muscle glove,” Production Engineering, vol. 11, no. 6, pp. 731–743, 2017. [122] S. Chatterjee, D. Hore, and A. Arora, “A wearable soft pneumatic finger glove with antagonistic actuators for finger rehabilitation,” in 2017 2nd International Conference on

Journal

Applied Bionics and BiomechanicsHindawi Publishing Corporation

Published: Aug 1, 2020

References