Design and Validation of Multichannel Wireless Wearable SEMG System for Real-Time Training Performance Monitoring

Serkan Örücü; Murat Selek

doi:10.1155/2019/4580645

Design and Validation of Multichannel Wireless Wearable SEMG System for Real-Time Training Performance Monitoring

Örücü, Serkan;Selek, Murat 2019-09-09 00:00:00 Hindawi Journal of Healthcare Engineering Volume 2019, Article ID 4580645, 15 pages https://doi.org/10.1155/2019/4580645 Research Article Design and Validation of Multichannel Wireless Wearable SEMG System for Real-Time Training Performance Monitoring 1 2 Serkan Oru¨cu¨ and Murat Selek Ermenek Vocational School, Karamanog˘lu Mehmetbey University, Karaman 70400, Turkey Vocational School of Technical Sciences, Konya Technical University, Konya 42130, Turkey Correspondence should be addressed to Serkan Oru¨cu¨; srknorucu@kmu.edu.tr Received 8 April 2019; Revised 20 July 2019; Accepted 16 August 2019; Published 9 September 2019 Guest Editor: Federica Verdini Copyright © 2019 Serkan Oru¨cu¨ and Murat Selek. ,is 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. Monitoring of training performance and physical activity has become indispensable these days for athletes. Wireless technologies have started to be widely used in the monitoring of muscle activation, in the sport performance of athletes, and in the examination of training eﬃciency. ,e monitorability of performance simultaneously in the process of training is especially a necessity for athletes at the beginner level to carry out healthy training in sports like weightlifting and bodybuilding. For this purpose, a new system consisting of 4 channel wireless wearable SEMG circuit and analysis software has been proposed to detect dynamic muscle contractions and to be used in real-time training performance monitoring and analysis. ,e analysis software, the Haar wavelet ﬁlter with threshold cutting, can provide performance analysis by using the methods of moving RMS and %MVC. ,e validity of the data obtained from the system was investigated and compared with a biomedical system. In this comparison, 90.95%± 3.35 for left biceps brachii (BB) and 90.75%± 3.75 for right BB were obtained. ,e output of the power and %MVC analysis of the system was tested during the training of the participants at the gym, and the training eﬃciency was measured as 96.87%± 2.74. contraction and relaxation of muscles. When academic 1. Introduction studies related to this subject are analysed, there are some In recent years, the monitoring of athlete performance has wearable biometric systems developed for the purpose of become indispensable for the health of athletes. Wireless the monitoring of performance during training. Some of technologies have started to be widely used in order to obtain these systems are intended for recording parameters like data for the purpose of examining training eﬃciency in the heart rate, respiration, location, and velocity or for esti- monitoring of muscle activation and sport performance of mating the levels of muscle fatigue [10–13]. Some of them athletes [1, 2]. It is possible to collect information about have been produced for the measurement of the SEMG athlete performance and rehabilitation, about preventing signals in laboratorial environment [14]. Another pro- muscle fatigue or injuries through posttraining analysis of portion of them has been designed for the purpose of perceiving dynamic muscle contraction during isolated SEMG signal obtained during the training [3–5]. Recording of SEMG signals in related muscles during training can be training through the SEMG [15]. ,e last proportion has extremely useful in increasing performance and preventing carried out low-cost experimental SEMG systems and disabilities [6]. matched the key features of the system with the existing Traits of SEMG signals obtained during training (fre- systems [16–18]. quency, severity, etc.) change depending on the muscle ,e most reliable method used in the adequacy and group measured and the severity of contraction [7–9]. In examination of muscle activation in physiological studies is these measurements, surface-type electrodes are used to the amplitude analysis carried out on SEMG signals, known determine and examine the activity of muscles during as MVC (maximum voluntary contraction) normalization 2 Journal of Healthcare Engineering like weightlifting and bodybuilding, for the performance [19]. Data with MVC normalization enable understanding of what capacity the muscle works, how eﬀective level evaluation of the athlete until the motor skills of the movement are improved and at necessary moments in muscles have reached through training and how much eﬀort a training requires from an athlete [20]. preventing the injury process by intervening in training. ,e simultaneous monitorability of athlete performance Based on these elements, a new wireless wearable SEMG during the process of training is a must for athletes at the data collection system has been introduced which enables beginner level to being able to carry out healthy training in performance monitoring and analysis at training time with sports like weightlifting and bodybuilding [21, 22]. ,is its real-time MVC normalization and contraction detection feature enables performance evaluation to be carried out processes. ,e SEMG circuit used in our system is designed momentarily during the time when there is no trainer or until by us to be used in future studies and to be developed the motor skills of the athlete concerning movement develop according to our needs. In the presented system, digital ﬁltering is also used in enough. A SEMG system, to be used during the training for this purpose [7, 23–25] has to addition to hardware ﬁltering in SEMG circuit. ,ese nu- merical ﬁlters are Haar wavelet ﬁlters with ,reshold cutting (i) Be able to provide the required SEMG data necessary based on (TCHW) and linear Kalman [37, 38]. Each nu- for monitoring training eﬃciency in performance merical ﬁltering method is tested together with hardware analysis ﬁltering. Results obtained from here will be determinative in (ii) Be able to ﬁlter the noise of movement during deciding the ﬁltering structure that can be used in future isotonic exercises and noise and distortions in stages of the system design. Subsequently, ﬁltered data are SEMG signals appearing as a result of other factors processed through moving RMS method containing the methods of moving average (MA) and root mean square (iii) Its procedures like calibration, etc., have to continue for a short time (RMS), scaled through MVC normalization, and a training support system that can carry out real-time performance (iv) ,e data obtained have to be at a close accuracy to analysis and monitoring. biomedical systems (v) Has to be simultaneously usable in a training 2. Materials and Methods environment For use in the industrial ﬁeld, various systems are 2.1. Isotonic Contraction. Isotonic contraction encompasses available for SEMG data collection and processing. To exercises where muscle tendons get shortened to generate investigate these, WB-EMG [26], BiometricsDatalog [27], movement. Any kind of movement, ranging from weight- Myo Armband [28], DelsysTrignio [29], BITalino [30], lifting to rowing and running, is in this category [39]. In Mbody3 [31], Mpower [32], MyoTrac [33], MyoWare [34], sport, an isotonic exercise is a training where the most Shimmer [35], and hospital [36] are such systems. ,e amount of strength is exerted on a particular muscle or systems speciﬁed in [26], [27], and [29] and the systems muscle group to increase that muscle mass or performance which we measure in hospital [36] are not wearable during in general. Due to the fact that human activity and athletic training. ,e system speciﬁed in [44] is wearable and performance necessitate these kinds of movement, isotonic supports wireless transmission but its production is exercises form the basis of a lot of training protocols [40]. It stopped. In terms of the electrodes used, and CMRR, there is possible to observe pathological changes or eﬃciency is no diﬀerence in all of these products and they comply obtained from the training through an examination of with the SENIAM criteria. ,e systems [26], [27], and [30] SEMG signals generated in muscles during these exercises do not have noise and data processing ﬁlters, and the [41]. systems in [28] and [29] use a Notch ﬁlter and a band-stop ﬁlter with narrow-bandwidth in hardware. ,e system in [26, 34] is designed for single-channel use but does not 2.2. SEMG Circuit Design. ,e SEMG circuit design details support multichannel monitoring. ,e systems in [31–34] are given below. ,e circuit consisting of 4 channels could are wearable and do not include contraction detection and monitor the biopotential change of 4 diﬀerent muscle groups simultaneous MVC analysis although they can monitor at the same time. So, it is possible to monitor biopotential multiple muscle groups. A summary of these comparisons changes occurring in muscles in symmetrical movements is presented in Table 1. that aﬀect multiple muscle groups (e.g., the Bench Press When the table is analysed, it is seen that all of these movement aﬀects pectoralis major and triceps muscles). ,e systems can simultaneously observe biopotential changes in circuit has in each channel, respectively, one in- muscle or muscle groups monitored during training, but none strumentation ampliﬁer, a inverting ampliﬁer, a low-pass of them include real-time MVC normalization and con- ﬁlter, a high-pass ﬁlter, and a full-wave rectiﬁer. ,e circuit traction detection procedures for performance analysis has a diode for input protection, a pointer indicating that the during training. circuit is working, and a start-up button. During working, ,at these features can be monitored simultaneously the LD1117 regulator was used for the Bluetooth feed and during the training process may be useful especially for the 7805 regulator for the +5 volt and − 5 volt op-amp feed beginner athletes to perform a healthy training in sports (Figure 1(a)). ,e SEMG signals we want to process are Journal of Healthcare Engineering 3 Table 1: Comparison of the SEMG acquisition systems. Real- Number ADC Signal Contraction time Connection System of Gain resolution Wearable Filter type CMRR type detection MVC type channels (bits) norm. Proposed SEMG 4 4400 12 Yes Hardware + software Yes Yes >90 Bluetooth system WB-EMG SEMG 1 100–10000 12 No No No No >90 Bluetooth Biometrics SEMG 8 1000 14 No No No No >90 Bluetooth datalog Myo SEMG 8 ≥1000 8 Yes Notch No No >90 Bluetooth armband Delsys SEMG 16 909 16 No Notch No No >90 RF Trignio BITalino SEMG Up to 6 1000 6–10 Yes No No No >90 Bluetooth Mbody3 SEMG Up to 6 ≥1000 24 Yes Hardware + software No No >90 Bluetooth Mpower SEMG 4 ≥1000 — Yes Hardware + software No No >90 Bluetooth MyoTrac SEMG 2 ≥1000 14 Yes Butterworth No No >90 Bluetooth MyoWare SEMG 1 ≥1000 — Yes No No No >90 Bluetooth Shimmer SEMG Up to 60 ≥1000 16 Yes Hardware + software No No >90 Bluetooth Hospital SEMG 8 1–10000 24 No Hardware + software Yes No >90 Usb MUAP signals whose amplitude is between 0 and 1.5 mVolt changed depending on the input signal parameters [48]. In (RMS). To process this electrical signal, it must ﬁrstly be these applications, analogue ﬁlters are used to eliminate noise from the signal in signal ampliﬁcation and processing ampliﬁed. In the system, this ampliﬁcation is done by in- creasing the diﬀerence between the two electrodes in bipolar circuits, to provide noise immunity, and to obtain the mode. While the obtained common signal is ampliﬁed in this necessary parts of the frequency band [49]. On the contrary, mode, the background noise is also suppressed. Two of the digital ﬁlters are used to ﬁlter signal residues named artifact probes activated from each channel are connected to the after motion and to analyse SEMG signal (feature extraction, circuit’s soil, like the reference probe [42] which is placed in time-frequency analysis, contraction detection, performance a more electrically remote area (preferably a neutral or close analysis, etc.) [41, 50]. to the bone region) while going to the ampliﬁer and ﬁlter In the circuit, analogue ﬁltering is performed by low- and circuits over INA 128P, which operates in a single diﬀer- high-pass ﬁlters. Ideal SEMG signals are observed between 50 Hz and 500 Hz and should be ﬁltered from frequency ential mode. In the ﬁrst step, ampliﬁcation was performed by using the INA 128P diﬀerential ampliﬁer (Figure 1(b)). components outside this range [51]. For this, the signal from the output of the instrumentation ampliﬁer is ﬁrst ﬁltered so As stated in [43], the reason why we use INA 128P is that the amplitude of the SEMG signal is low and that the that the gain is 1 in the high-pass ﬁlter (HPF) using TL072 ampliﬁer to be used due to other factors like noise must have with a cutoﬀ frequency of about 48 Hz (Figure 1(d)). ,e a high input impedance and a high common mode rejection components of the EMG signal above 500 Hz are ﬁltered rate (CMRR> 95 dB). ,is ampliﬁer has the required fea- through a 2nd order Sallen–Key low-pass ﬁlter (LPF) using tures with CMRR >120 dB and 10 GΩ input impedance. TL072. ,rough this section, resistance and capacitor values When we set the gain value for the 60 Hz input signal to are designed so that the cutoﬀ frequency is approximately G � 74.52 using INA 128P in our system, approximately 482 Hz, the quality factor is 0.5, and the gain is 1 108 dB CMRR was obtained as stated in the technical (Figure 1(e)). ,e reason we prefer the Sallen–Key topology we use in the circuit is that this ﬁlter has the ability to document in [44]. ,e reason for selecting a 60 Hz input signal in the system design is that the SEMG signal is produce a quadratic low-pass reaction with better selectivity (higher pole) and various possible approaches (Butterworth, dominant in the range of 50 Hz to 150 Hz. To obtain a processable signal amplitude in the second stage, TL072 was Chebyshev, ,omson-Bessel, etc.) [43, 47, 49]. ,is will help used as shown in Figure 1(c) as an active inverting ampliﬁer. us in our future work. At this stage was the ampliﬁer gain approximately G � 59 ,en, the whole SEMG signal was moved to the positive and the CMRR approximately 100 dB by using the 60 Hz level using the full-wave rectiﬁer (Figure 1(f)). With this input signal as stated in [45]. process, it is possible to analyse the low-frequency oscilla- In SEMG applications, analogue (hardware) and digital tions by overcoming the high-pass nature of the SEMG (software) ﬁlters are used to remove unwanted component signal [52]. ,us, it is aimed to use the circuit except for the noises and process the necessary parts in the SEMG signal training eﬃciency, also in the ﬁelds of prosthesis control and ergonomics. [46]. Analogue ﬁlters remove anything above or below a selected cut frequency, while digital ﬁlters make this process ,e Pic16F1786 microcontroller with connected full-wave rectiﬁer outputs contains 11 12 bit A/D (Analogue/Digital) more precise as they can be programmed [47]. ,is certainty is due to the fact that the features of digital ﬁlters can be converters. ,e data obtained from the rectiﬁer of each channel 4 Journal of Healthcare Engineering U2 U7 7805 LD1117V33 D1 U3 1 3 3 2 StereoJack R1 VI VO IN OUT J1 1N4007 680 2 inaOut GND U1 C2 C4 C11 C8 C9 RG1 C10 1 6 7805 220uF 220uF 100uF 1 100uF RG2 J2 2 13 5 2 V1 100uF 100uF VO REF C3 INA122 C1 POWER 220uF 220uF –5V (a) (b) C6 R3 R5 330k 33k 1nF U4:B U5:A U4:A R7 R6 + + C5 3 7 1 AMPOUT R4 + HPFOUT 330k 330k 2 1 6 R2 – INAOUT – hpfOut LPFOUT AMPOUT 33k 100nf C7 5.6k 1nF TL072 TL072 TL072 (c) (d) (e) R8 U8 10k RECTOUT RA0 RC0 Tx 3 12 U6:A Rx U5:B RA1 RC1 4 13 D2 RA2 RC2 1N4148 3 RA3 RC3 +3.3V 6 15 + 1 Vcc RA4 RC4 GND 7 16 R10 2 GND 6 – HC06 RA5 RC5 – RectOut TX Tx 10 17 RA6 RC6 Rx RX 10k 9 18 LPFOUT RA7 RC7 D3 TL072 TL072 21 25 1N4148 RB0 RB4 R11 R9 R12 RB1 RB5 23 27 RB2 RB6/ICSPCLK 24 28 RB7/ICSPDAT RB3 10k 10k 10k RE3/MCLR/VPP PIC16F1786 (f ) (h) (g) fe d c b In. (i) Figure 1: Block diagram and mounted state of the SEMG circuit. (a) Regulator circuit. (b) Instrumentation ampliﬁer. (c) Inverting st nd ampliﬁer. (d) 1 -order HPF. (e) 2 -order Sallen–Key LPF. (f) Full-wave rectiﬁer. (g) PIC 16F1786. (h) Bluetooth module. (i) Mounted state of the SEMG circuit. in the circuit are connected, respectively, to the RA0-RA3 received and processed by the data collection program inputs of this controller. ,is microcontroller performs the written in the C# language. ,e digitalized SEMG data in A/D conversion in 20 ms time intervals through the pro- the data collection program are processed through digital gram we write. ,e converted channel data are turned into ﬁlters. ,e PCB (printed circuit board) of the circuit is a string, and this sends data from the RC0 output to the designed to be 10 cm × 10 cm in size, and as stated in [53], the Bluetooth module (Figure 1(g)). ,e transmitted data have PCB tracks are intended to be exposed to as little noise as a resolution of 2.4 μV in each step. Data sent at 4800 bps possible. ,e mounted state of the circuit shown in Figure 1(i) speed via the HC-06 Bluetooth module (Figure 1(h)) are is boxed and placed inside a wearable belt. ,e necessary –5V +5V 4 8 –5V 2 4 8 GND +5V GND +5V –5V 4 8 +5V 4 8 –5V +5V +3.3V 4 8 –5V +5V –5V +5V Journal of Healthcare Engineering 5 Table 2: Information about age, gender, weight, and height of the energy for the operation of the circuit was obtained from subjects. 1000 mAh lithium batteries. It is intended to minimize power line interference (PLI) without the need for any insulation, as Participant no. Age Gender Weight (kg) Height (cm) stated in [54] using battery in the system. 1 21 Male 80 163 2 25 Male 82.3 178 3 29 Male 87 180 2.3. Participants and Setup. Five males and two females 4 33 Male 85 177 voluntarily participated in our study and have at least 2 years 5 37 Male 104.6 193 of experience in strength training. ,e information of the 6 24 Female 70 180 participants is shown in Table 2. 7 27 Female 68 172 ,e participants were informed about the content of our study, and a signed consent form was obtained from all of them. All exercises and measurements were made under the supervision of a specialized trainer. As described in the recommendations of the European initiative known as SENIAM (surface electromyography for noninvasive muscle evaluation of muscles) by selecting 10 mm diameter electrodes shown in Figure 2 for SEMG, the bipolar con- ﬁguration is located 1–2 cm away from the centre of the muscle and the reference electrode is placed in a region that is electrically neutral according to the action [51]. ,e connection between the electrodes and the circuit channels is provided by using armoured cables which have 3.5 mm Figure 2: Example view of electrodes and shielded cables. ends, 3 colour code (red, green, and blue) and labelled contacts (L, F, and R), as shown in Figure 2. When the system is modelled, it was aimed to minimize the Our experiments consist of 3 parts. In the ﬁrst part, 8 distortions in data by estimating the k parameter speciﬁed by repetitions and 1 set of alternate dumbbell curl (ADBC) x in SEMG data array at a particular time as X : training was performed using a maximum load of 60–70%. 􏽢 􏽢 X � K · Z + (1− )K · X . (1) In this section, ﬁrstly, it is investigated whether the ana- k k k k k− 1 logue ﬁlter data obtained from the circuit in the training Here, Z expresses the measuring data wanted to be reﬂect the biopotential activity changes that occur during absolutized, K the Kalman gain and X the measuring k k− 1 the training. In the sequel, the analogue ﬁlter data obtained data belonging to the previous stage. If the system is from the circuit are processed by means of Kalman and modelled through this information, a model consisting of threshold cut Haar wavelet ﬁlter (TCHW) to eliminate calculation (2) and update (3) is obtained. noise sources and to investigate the perceptibility of the x � Ax + Bu + w , (2) isotonic contractions. k k− 1 k k− 1 In the second part, the accuracy of the developed system was compared with the biomedical system (Viking on z � Hx + v . (3) k k k Nicolet EDX) used in Karaman State Hospital (See Table 1). In (2), any x is expressed as a linear combination of the In this comparison, the RMS values obtained from both k next control signal k of its previous value and the noise of the systems were used. process. In (3), any measurement value making certain of the In the third part, the availability of moving RMS and accuracy of which we are not sure is accepted to be a linear %MVC values as the screen output of the system was combination of the signal value and the noise of the investigated in terms of performance feedback. For this measurement. purpose, ﬁrst, the moving RMS values obtained by asking In HW, the main wavelet acts as the wavelet transform users to perform a second ADBC (8 repetitions 1 set) but is scaled and shifted during this procedure of wavelet training were recorded. In addition, a %MCV mea- transform [35]. Scaling corresponds to the widening and surement was made by asking all users in the training constriction of the signal (f(t)) and the shift to the wave environment to lift 5 kg dumbbell and maximum weight shift (τ) in the timescale axis (t) in the following equation (Men 17.5 kg, 20 kg, and 25 kg dumbbell; women 12.5 kg [57, 58]: and 15 kg dumbbell) they can. − jωt F(ω, τ) � 􏽚 f(t)w(t − τ)e dt. (4) 2.4. Kalman and TCHW Filters. Kalman ﬁlter is used to estimate the system status from input and output in- formation with the previous information of a model in a HW is a wavelet-based, scaled, “square-shaped” array of dynamic system indicated by the state-space model [55, 56]. functions. ψ(t), the main function of HW (5), and also φ(t), 6 Journal of Healthcare Engineering 􏽳�� a scaling function (6), are deﬁned in t time interval given as (11) follows: f � [f(t)] dt, rms T − T 2 1 1 ⎧ ⎪ ⎪ 1, 0≤ t≤ , ⎪ 􏽳�� ⎪ 2 ⎨ (12) f � lim 􏽚 [f(t)] dt. rms T⟶∞ ψ(t) � (5) T 0 − 1, < t≤ 1, Another method we use as MA is the technique of analysing changes in a data set to estimate long-term trends. 0, otherwise, For a given N time window, if the values s , s , s ,. . ., s 1 2 3 n corresponding to this time interval of the S variable shown in ⎧ ⎪ 1, 0≤ t≤ , ⎨ the times t , t , t , . . ., t are known, the MA window size is 1 2 3 n φ(t) � (6) ⎪ deﬁned as N � 2k + 1 and processed as speciﬁed in 0, otherwise, +k MA � s . ,e Haar function ψ is deﬁned as shown in 􏽘 (13) n,k i− j j�− k n/2 n ψ (t) � 2 ψ 2 t − k , t ∈ R. (7) n,k ,us, changes in the time window given at the j moment are obtained by averaging the time series of the k Since the SEMG signals are user-based, SEMG signals between isotonic muscle contractions may vary according to time in the j moment. Instead of using the above- mentioned RMS and MA methods separately, the moving the individual. In the method we use with HW, the indi- vidual waits for approximately 2–4 seconds with the weight RMS method was used in our system by calculating the RMS value in a moving window, which is a combination of in his hand before starting training and in the meantime, the procedure of threshold cutting in the system can be carried these methods. In this method, the operation can be performed at any t time interval of the moving window; out. ,e threshold cutting is based on the calculation of the average value (8), the standard deviation (9), and the signal therefore, it acts as a ﬁlter in a certain time interval, as slope (10): shown in (14). In this way, the processing of the data obtained according to the variable speed of the replays in A � ∗ 􏽘 x , (8) i the training sets gets easier. In this equation, n refers to the i�1 length of the window, while x(k) refers to the data within 􏽶�� the window: 1/2 (9) σ � 􏽘 x − μ􏼁 , i ⎝ ⎠ ⎛ ⎞ i�1 (14) x [i] � 􏽘 x [k] . RMS j�(i− N+1) 􏽐(x − x)(y − y) So, it can be measured how much power is obtained from s � . (10) 􏽐(x − x) the muscle through the moving RMS value. ,e MVC (maximum voluntary contraction-maximum Here, x is the value added to the average, μ is the average amplitude of the signal) normalization is widely used in value and N is the number of the total value. After the values SEMG signals as an amplitude analysis technique. ,e re- of the average, standard deviation and slope are calculated sults are shown as a percentage (%MVC) of the MVC value and all SEMG signals complying with this condition are that can be used to create a common background when equalled to zero. ,us, the signals between the voluntary comparing data between subjects [60, 61]. SEMG signals contractions can be eliminated. depend on the user and have a structure that can cause records to change even when measured from the same 2.5. RMS, MA, and %MVC. After the SEMG signal is cap- position with the same motion. ,erefore, MVC normali- tured, the commonly used RMS or MA values are analysed zation is used to eliminate this diﬀerence and to enable data by using [59]. In RMS analysis, the SEMG signal is subjected comparison between subjects [61]. MVC expresses the to a set of mathematical operations designed to measure the highest value obtained in a repeat during this measurement power of change. ,us, the intensity and duration of events to normalize SEMG signals obtained for a speciﬁc purpose, like muscle contractions can be investigated. ,erefore, the while SMVC (submaximal voluntary contraction) refers to RMS value is a parameter chosen during contraction and the voluntarily recorded SEMG data. %MVC corresponds to reﬂects the level of physiological activity in the body. the multiplication of the normalized value of according to Mathematically, the RMS value of a continuous-time SMVC’s MVC with 100 [62, 63]: waveform is the square root of a function deﬁning the SMVC continuous waveform shown in f (t) in the following, de- %MVC � (15) 􏼒 􏼓∗ 100. ﬁned in the range T ≤ t≤ T : MVC 1 2 Journal of Healthcare Engineering 7 (a) (d) MVC analysis 4x Surface EMG inst. amp. Amplifier 1-order HPF EMG electrode G = 74.5 G = 59 M-RMS calculation Haar wavelet filter 2-order 12 bit A/D Bluetooth Sallen–Key LPF converter module PC (b) (c) Figure 3: Overview of the system. (a) Connecting electrodes before training (Photoshoot by Orucu). (b) Block diagram of the SEMG circuit. (c) Block diagram of the analysis software. (d) User interface of the analysis software. ,us, it can be scaled how much power is obtained from the training speed (Figure 3(c)). After this process, the the muscle or muscle groups investigated in repetitions in SMVC value of each repetition in each set of the training is processed according to the previously saved MVC values. each set of training. ,en, %MVC values are displayed on the screen in separate graphs according to the channels from which the data are 2.6. Proposed System. Our system has the ability to follow taken. Finally, they are saved to the database in “.csv,” the biopotential changes of four diﬀerent superﬁcial muscle “.dat,” and “.xlsx” formats (Figure 3(d)). groups at the same time. ,e reason why the system is designed with 4 channels is that most movements used in 3. Results and Discussion bodybuilding and weight training activate at least 1 or 3 muscle groups at the same time. ,e system takes the 3.1. Analogue + Digital Filtered Data from the System. ,e biopotential signals of the muscles that are activated during analogue-ﬁltered data of the ﬁrst 4 repetitions of ADBC training through surface electrodes (Figure 3(a)), and then, training performed by participant number two is shown in ﬁrst it ampliﬁes them in the instrumentation and ampliﬁer Figure 4(a), marked as 4(a) and 4(b) for each repetition. st parts in the SEMG circuit, after it ﬁlters them with the 1 - ,e left BB (LBB-Left Biceps Brachii) data are obtained nd degree high pass and 2 -degree Sallen–Key low-pass ﬁl- from CH1 (ﬁrst channel of the SEMG circuit), and the right ters. ,ese analogue-ﬁltered signals are sent to the com- BB (RBB-Right Biceps Brachii) data are obtained from CH2 puter via Bluetooth after a 12 bit analogue-to-digital (the second channel of the SEMG circuit). From the data conversion (Figure 3(b)). By the software we developed in obtained, some decrease in Rep2b, Rep3a, Rep3b, and Rep4a C# language, all SEMG channel data received by the (between 100 and 200 μV) and a data change during pushing computer are digitally ﬁltered and then they calculated the the weight down (relaxation period of the muscle) in Rep 4b moving RMS values in time windows that vary according to were observed. As we consulted with the professor of 8 Journal of Healthcare Engineering Rep2a Rep1b Rep4a Rep3b Rep1a Rep3a Rep2b Rep4b Time (millisecond) CH1-left biceps brachii CH2-right biceps brachii (a) Rep1 Rep2 Rep3 Rep4 Rep5 Rep6 Rep7 Rep8 Rep9 Rep10 Rep11 Rep12 Time (millisecond) CH3 CH4 (b) Figure 4: Sample analogue ﬁltered data obtained from the SEMG circuit during training: (a) Sample results of participant number two, (b) sample results of participant number six. ˘ both relax normally in Rep10, in which distortion in Physical Education and Sports Teaching (Karamanoglu Mehmetbey University), he stated that the fall was caused by movement appears as a result of fatigue in Rep11 and the distortion of movement. According to the consultant Rep12. In addition, the data of other participants obtained professor, this change appeared to have been caused by the from these trainings are presented in Figure 5. prolongation of the activation period of the muscle as a In Figure 6, the data, processed with TCHW and Kalman result of pushing the weight down more slowly as speciﬁed ﬁlters, of two repetitions in training, belonging to the right in [64, 65]. BB muscle, conducted by the participant numbered 4, are Other data of training performed by participant shown. In this Figure, 6(a) shows the analogue ﬁltered state number four are shown in Figure 4(b). In this training, LBB of the SEMG signal, and 6(b) shows the preliminary mea- data were obtained from CH3 (the third channel of the surement of the threshold cut-out. ,e average and standard SEMG circuit) and RBB data were obtained from CH4 (the deviation measured here were found as 61.11± 51.61 μV, and the slope was found as 0.005⁰. ,e signal ﬁltered with TCHW fourth channel of the SEMG circuit). When the results obtained are investigated in accordance with contraction after this procedure is shown in 6(c), and the signal pro- and relaxation situations as speciﬁed in [65, 66] which cessed through Kalman ﬁlter is shown in 6(d). Filtering consultant professor pointed, it is observed that BB muscles results indicate that the TCHW method produces better contract and relax normally in Rep1, Rep5, Rep7, and Rep9 results in ﬁltering unwanted signals and contraction de- and BB muscles contract fast and relax normally in Rep2. It tection compared to the method of Kalman ﬁlter. As a result is observed that the left BB contracts more than the right BB of these processes, it was decided to use TCHW ﬁlter in our does and both relax normally in Rep3, that the required system. support is taken from other regions and movement is ruined in Rep4 and that the left BB muscle contracts more, the right BB muscle contracts normally and both relax 3.2. Comparison Results with the Existing Biomedical System. ,e accuracy of the data obtained from our system was normally in Rep6 and Rep8. It is observed that the left BB compared through the data belonging to two men and two contracts normally and the right BB contracts more and SEMG signal (μVolt) SEMG signal (μVolt) 1009 Journal of Healthcare Engineering 9 Time (millisecond) CH1-left biceps CH3-right biceps (a) Time (millisecond) CH2-left biceps CH4-right biceps (b) Time (millisecond) CH1-right biceps CH4-left biceps (c) Figure 5: Continued. SEMG signal (μVolt) SEMG signal (μVolt) SEMG signal (μVolt) 1 1 9 9 17 17 33 36 37 41 56 53 61 65 65 69 81 77 81 91 109 117 113 121 117 125 136 129 129 146 133 151 161 153 10 Journal of Healthcare Engineering Time (millisecond) CH1-left biceps CH4-right biceps (d) Time (millisecond) CH2-left biceps CH3-right biceps (e) Figure 5: Data of other participants obtained from these trainings. (a) Results of participant number one. (b) Results of participant number three. (c) Results of participant number ﬁve. (d) Results of participant number six. (e) Results of participant number seven. 1200 250 CH3-left BB Raw data 0 0 0 2000 4000 6000 8000 0 10002000300040005000 t as millisecond t as millisecond (a) (b) Figure 6: Continued. Mean CH3-left BB as μV raw SEMG signal (μVolt) SEMG signal (μVolt) SEMG noise as uV 39 51 53 53 55 55 57 57 59 81 Journal of Healthcare Engineering 11 1200 525 CH3-left BB 400 110 0 2000 4000 6000 8000 t as millisecond t as millisecond (c) (d) Figure 6: Comparison of the ﬁltering results. (a) SEMG data without the ﬁlter. (b) Premeasurement for threshold ﬁlter. (c) SEMG signal with threshold + HW ﬁlter. (d) SEMG signal with Kalman ﬁlter. (a) (b) Figure 7: (a) A measurement taken in the hospital environment and a photograph of the current biomedical system. (b) A photograph taken at the gym before training. Table 3: Moving RMS Results in Gym and Hospital. Note that “M” denotes the measurement number; “BB” denotes biceps brachii; “S” denotes system; “H” denotes hospital, “MN” denotes muscle name. Participants/weight (no./kg) M MN Type 1/idle 2/idle 3/idle 4/idle 1/5 2/5 3/5 4/5 1/25 2/25 3/15 4/12.5 S 70.69 69.72 51.18 43.82 123.69 129.54 97.54 93.64 914.7 935.98 566.98 547.64 Left BB H 67.13 72.31 47.24 45.9 137.42 141.94 108.66 101.05 950.94 1112.53 616.53 604.36 S 71.4 69.75 49.66 42.45 119.11 127.41 96.86 93.95 960.71 937.69 565.69 515.43 Right BB H 69.64 70.51 50.22 43.93 135.57 143.13 107.93 97.14 943.82 1117.15 615.15 545.64 S 69.86 68.84 52.03 43.15 121.82 128.9 95.71 93.8 907.35 934.5 563.5 518.06 Left BB H 70.39 69.61 51.76 43.75 138.87 139.69 105.78 101.55 942.14 1116.89 614.89 595.59 II S 69.84 71.34 49.01 46.68 122.96 126.95 96.5 93.61 950.6 932.61 562.61 526.48 Right BB H 71.82 67.83 50.25 43.27 136.52 142.8 106.73 102.46 1002.4 1110.94 612.94 598.81 S 69.45 70.89 50.96 46.22 124.61 128.91 97.88 89.07 907.15 934.34 564.34 511.19 Left BB H 68.57 69.97 50.24 46.71 138.81 139.69 106.74 105.67 1000.8 1115.11 614.11 583.55 III S 71.64 67.65 52.01 43.55 122.97 129.74 95.38 92.52 948.63 930.36 562.36 539.49 Right BB H 68.56 69.63 49.72 44.79 135.94 145.8 107.5 104.28 1000.9 1110.61 612.61 543.35 women with the SEMG device in Karaman State Hospital times with breaks of 90 seconds. In this procedure, ﬁrst the (Figure 7). data given from the hospital system were recorded and then As shown in Table 3, this procedure was carried out the moving RMS was calculated on the analogue and digital through the data of 108 measurements in total, obtained ﬁlter data obtained from the system. through volunteers being unattached, lifting dumbbells of In the system designed as a result of this measurement, 5 kg and the maximum weight they could lift isometrically accuracies of 90.95%± 3.35 for the left BB and 90.75%± 3.75 (1 RM) ﬁrst in the gym, then in the hospital system for three for the right BB were obtained. CH3-left BB as μV with threshold Haar CH3-left BB as μV with Kalman filter 1,000 1,500 2,000 2,500 3,000 3,500 4,000 4,500 5,000 5,500 6,000 6,500 7,000 12 Journal of Healthcare Engineering Table 4: Moving RMS results in gym as training feedback. Muscles and participants Rep1 Rep2 Rep3 Rep4 Rep5 Rep6 Rep7 Rep8 LBB 1 862 798 738 683 782 556 715 741 LBB 2 845 779 852 786 590 812 796 766 LBB 3 757 725 721 560 712 699 645 736 LBB 4 810 841 840 804 828 832 791 830 LBB 5 704 802 651 670 604 354 558 701 LBB 6 387 413 395 354 367 403 381 370 LBB 7 316 328 372 346 377 302 328 319 RBB 1 876 833 811 790 815 846 704 653 RBB 2 823 817 847 834 649 747 621 770 RBB 3 821 793 766 696 566 884 685 785 RBB 4 832 853 856 821 819 808 809 815 RBB 5 815 763 750 753 718 707 725 714 RBB 6 389 422 418 350 371 402 361 378 RBB 7 331 380 365 351 372 348 314 341 1000 1000 900 900 800 800 700 700 600 600 500 500 400 400 300 300 200 200 100 100 0 0 LBB LBB LBB LBB LBB LBB LBB RBB RBB RBB RBB RBB RBB RBB 1 2 3 4 5 6 7 1 2 3 4 5 6 7 Muscles and participants Rep1 Max. rep Min. rep Last rep Figure 8: ADBC results of participants. 3.3. Moving RMS and %MVC Values. During the training, accuracy. As digitally ﬁltered data are compared, it is the volunteers were asked to perform a second training in seen that TCHW method produces better results com- order to obtain the moving RMS values given back to the pared to Kalman ﬁlter. TCHW can soften data as pro- user as feedback. ,e results are presented in Figure 8 and cessable and can also completely ﬁlter out unwanted Table 4 in terms of ease of investigation. signals between muscle contractions. It also eliminates ,us, it can be seen that the system can achieve mini- the distortions in data expressed as artifact. Kalman ﬁlter mum and maximum values of biopotential changes in appears to soften the data but not to be able to completely muscles during training as in [66, 67]. ﬁlter the signal between muscle contractions. Moreover, it is seen that the system can scale the strength obtained Finally, the users were asked to lift 5 kg of dumbbell and the maximum weight they could lift. ,us, the %MVC was as moving RMS during the training on the basis of % MVC with the success rate of 96.87%± 2.74 in terms of measured to be used in performance feedback through the obtained moving RMS values. ,e results obtained are eﬃciency. ,is allows the data obtained to be used in the presented in Table 5. simultaneous performance monitoring and analysis of If Table 5 is analysed, it can be seen that the system can athletes. measure eﬃciency during training with the success rate of 96.87%± 2.74 based on %MVC. 4. Conclusion When data obtained from the designed SEMG system are compared with data obtained from the systems used ,anks to this system, it is thought that athletes will be in the biomedical ﬁeld, it is seen that it has 90.85% able to examine their performances instantly for each M-RMS values (μV) Journal of Healthcare Engineering 13 Table 5: %MVC results in gym. Acknowledgments Muscle name Part. no. kg SMVC (μV) MVC (μV) ,e authors would like to thank Assistant Professor Dr. MVC Yusuf Er (Karamanoglu ˘ Mehmetbey University Physical 5 138.3 850.26 16.26 1 Education and Sports Teaching-Recreation Management) 17.5 845.47 850.26 99.43 for his helpful advice on various technical issues and Atilla 5 138.30 992.6 13.93 Sonmezı ¨ ¸sık (Antalya Sport Center) for his training support. 25 987.15 992.6 99.45 5 152.6 960.13 15,89 20 898.25 960.13 93,55 References 5 155.4 963.30 16.13 LBB 4 20 929.1 963.30 96.44 [1] R. M. Howard, R. Conway, and A. J. Harrison, “Muscle ac- 5 158.9 967.5 16.42 tivity in sprinting: a review,” Sports Biomechanics, vol. 17, 20 956.76 967.5 98.88 no. 1, pp. 1–17, 2018. 5 89.5 615.7 14.53 [2] D. R. Clark, M. I. Lambert, and A. M. Hunter, “Contemporary 15 614.74 615.7 99.84 perspectives of core stability training for dynamic athletic 5 93.5 545.2 17.14 performance: a survey of athletes, coaches, sports science and 12.5 538.9 545.2 98.84 sports medicine practitioners,” Sports Medicine-Open, vol. 4, 5 136.5 875.9 15.58 no. 1, pp. 32–42, 2018. 17.5 854.15 875.9 97.51 [3] J. Hussain, K. Sundaraj, Y. F. Low, L. C. Kiang, S. Sundaraj, 5 140.22 996.4 14.07 and M. A. Ali, “A systematic review on fatigue analysis in 25 985.92 996.4 98.94 triceps brachii using surface electromyography,” Biomedical 5 154.42 972.4 15.88 Signal Processing and Control, vol. 40, pp. 396–414, 2018. 20 893.56 972.4 91.89 [4] R. Lovell, M. Knox, M. Weston, J. C. Siegler, S. Brennan, and 5 156.8 968.7 16.18 P. W. M. Marshall, “Hamstring injury prevention in soccer: RBB 4 20 930.76 968.7 96.08 before or after training?,” Scandinavian Journal of Medicine 5 164.45 972.65 16.90 and Science in Sports, vol. 28, no. 2, pp. 658–666, 2018. 20 910.5 972.65 93.61 [5] S. Skals, J. Vinstrup, E. Sundstrup, M. D. Jakobsen, 5 97.1 614.4 15.80 C. H. Andersen, and L. L. Andersen, “Shoulder and arm 15 609.9 614.4 99.26 muscle activity during elastic band exercises performed in a 5 92.2 515.16 17.89 hospital bed,” De Physician and Sportsmedicine, vol. 46, no. 2, 12.5 506.3 515.16 98.28 pp. 233–241, 2018. [6] M. Cardinale and M. C. Varley, “Wearable training-moni- toring technology: applications, challenges, and opportuni- training and make their training more eﬃcient. It is ties,” International Journal of Sports Physiology and Performance, vol. 12, no. s2, pp. S2–S55, 2017. possible to create intelligent training corners by using the [7] A. D. Vigotsky, I. Halperin, G. J. Lehman, G. S. Trajano, and system in gyms. It is thought that the system can easily be T. M. Vieira, “Interpreting signal amplitudes in surface used by athletes, trainers, kinesiologists, and re- electromyography studies in sport and rehabilitation sci- habilitation experts in bodybuilding trainings and re- ences,” Frontiers in Physiology, vol. 8, pp. 1–15, 2018. habilitation processes. It is possible to improve system [8] E. Zemkova´ and D. Hamar, “Sport-speciﬁc assessment of the features by increasing the number of channels, further eﬀectiveness of neuromuscular training in young athletes,” reducing the PCB size and adding extra sensor. It can be Frontiers in Physiology, vol. 9, pp. 1–27, 2018. possible to follow more complicated movements (deadlift, [9] M. Vromans and P. Faghri, “Electrical stimulation frequency barbell row, etc.) by increasing the number of channels. By and skeletal muscle characteristics: eﬀects on force and fa- making the size of system smaller, it can be possible to tigue,” European Journal of Translational Myology, vol. 27, place it into textile product. In addition, by adding the no. 4, pp. 239–245, 2017. [10] M. Kos and I. Kramberger, “A wearable device and system for pulse oximetry sensor to the system, oxygen consumption movement and biometric data acquisition for sports appli- can be observed during the training. In our future studies, it cations,” IEEE Access, vol. 5, pp. 6411–6420, 2017. is being thought of supporting the system with an image [11] O. Ozkaraca, A. Hakan Isik, and I. Guler, “Biological signals processing system in order to determine movement dis- sensing with wearable technologies and an ECG athlete de- tortions in addition to use it for monitoring training sign,” Recent Patents on Biomedical Engineeringe, vol. 4, no. 2, performance and eﬃciency. pp. 74–83, 2011. [12] R. T. Li, S. R. Kling, M. J. Salata, S. A. Cupp, J. Sheehan, and Data Availability J. E. Voos, “Wearable performance devices in sports medi- cine,” Sports Health: A Multidisciplinary Approach, vol. 8, ,e data used to support the ﬁndings of this study are in- no. 1, pp. 74–78, 2015. cluded within the article. [13] S. Saponara, “Wearable biometric performance measurement system for combat sports,” IEEE Transactions on In- strumentation and Measurement, vol. 66, no. 10, pp. 2545– Conflicts of Interest 2555, 2017. ,e authors declare that there are no conﬂicts of interest [14] I. S. Dhindsa, R. Agarwal, and H. S. Ryait, “Performance regarding the publication of this paper. evaluation of various classiﬁers for predicting knee angle from 14 Journal of Healthcare Engineering electromyography signals,” Expert Systems, vol. 36, no. 3, [29] Delsys, Natick, MA, USA, “Trigno lab,” 2017, http://www. pp. e12381–14, 2019. delsys.com/products/wireless-emg/trigno-lab/. [15] Y. Yin, Y. Zeng, X. Chen, and Y. Fan, “,e internet of things in [30] bitalino, Lisbon, Portugal, “EMG 151015,” 2017, http:// healthcare: an overview,” Journal of Industrial Information bitalino.com/datasheets/EMG_Sensor_Datasheet.pdf/. Integration, vol. 1, pp. 3–13, 2016. [31] Muscle Monitor, Kuopio, FINLAND, “myontec,” 2019, [16] V. Bonaiuto, N. Lanotte, C. Romagnoli, N. Silvaggi, and https://www.myontec.com/musclemonitor/manuals/Old/3_ G. Annino, “Design of a wireless wearable DAQ system for the 1/Muscle%20Monitor%20-%20User%20Guide.pdf12. evaluation of sports performances,” Proceedings, vol. 2, no. 6, [32] Mpower, Vimpeli, Finland, “Mpower,” 2019, http://mpower- pp. 290–295, 2018. bestrong.com/manual/MPOWER_user_manual.pdf. [17] S. Heywood, Y. H. Pua, J. McClelland et al., “Low-cost [33] MyoTrac, Montreal, Quebec, Canada, “MyoTrac3,” 2019, electromyography—validation against a commercial system http://www.thoughttechnology.com/pdf/mar870-00% using both manual and automated activation timing 20myotrac3%20rehab.pdf. thresholds,” Journal of Electromyography and Kinesiology, [34] H. P. Da Silva, “Physiological sensing now open to the world: vol. 42, pp. 74–80, 2018. new resources are allowing us to learn, experiment, and create [18] G. Yang, J. Deng, G. Pang et al., “An IoT-enabled stroke imaginative solutions for biomedical applications,” IEEE rehabilitation system based on smart wearable armband and Pulse, vol. 9, no. 2, pp. 9–11, 2018. machine learning,” IEEE Journal of Translational Engineering [35] A. Burns, B. R. Greene, M. J. McGrath et al., in Health and Medicine, vol. 6, pp. 1–10, 2018. “SHIMMER —A wireless sensor platform for noninvasive [19] A. Ranavolo, G. Chini, A. Silvetti, S. Mari, M. Serrao, and biomedical research,” IEEE Sensors Journal, vol. 10, no. 9, F. Draicchio, “Myoelectric manifestation of muscle fatigue in pp. 1527–1534, 2010. repetitive work detected by means of miniaturized sEMG [36] Viking On Nicolet, Planegg, Germany, “Viking on nicolet,” sensors,” International Journal of Occupational Safety and 2019, https://www.neuroswiss.ch/view/data/5962/Nicolet- Ergonomics, vol. 24, no. 3, pp. 464–474, 2018. Viking-EDX.pdf. [20] A. Cisnal, R. Alonso, J. P. Turiel, J. C. Fraile, V. Lobo, and [37] K. Biron, K. Englehart, and P. Parker, “A study of kalman V. Moreno, “EMG based bio-cooperative direct force control of ﬁltering applied to myoelectric signal state tracking,” CMBES, an exoskeleton for hand rehabilitation: a preliminary study,” in vol. 35, pp. 1–4, 2012. Converging Clinical and Engineering Research on Neuro- [38] T. Boaventura, L. Hammer, and J. Buchli, “Interaction force rehabilitation III. ICNR 2018. Biosystems & Biorobotics, estimation for transparency control on wearable robots using L. Masia, S. Micera, M. Akay, and J. Pons, Eds., vol. 21, Springer, a kalman ﬁlter,” in Converging Clinical and Engineering Re- Cham, Switzerland, 2019. search on Neurorehabilitation II. Biosystems & Biorobotics, [21] T. D. Williams, D. V. Tolusso, M. V. Fedewa, and M. R. Esco, J. Ibañez, ´ J. Gonzalez-Vargas, ´ J. Azor´ın, M. Akay, and J. Pons, “Comparison of periodized and non-periodized resistance Eds., vol. 15, Springer, Cham, Switzerland, 2017. training on maximal strength: a meta-analysis,” Sports [39] B. J. Schoenfeld, “,e mechanisms of muscle hypertrophy and Medicine, vol. 47, no. 10, pp. 2083–2100, 2017. their application to resistance training,” Journal of Strength [22] R. A. Panariello, T. J. Stump, and F. A. Cordasco, “,e lower and Conditioning Research, vol. 24, no. 10, pp. 2857–2872, extremity athlete: postrehabilitation performance and injury prevention training,” Operative Techniques in Sports Medi- [40] B. J. Wallace, H. C. Bergstrom, and T. A. Butterﬁeld, cine, vol. 25, no. 3, pp. 231–240, 2017. “Muscular bases and mechanisms of variable resistance [23] J. Gondin, P. J. Cozzone, and D. Bendahan, “Is high-frequency training eﬃcacy,” International Journal of Sports Science and neuromuscular electrical stimulation a suitable tool for Coaching, vol. 13, no. 6, pp. 1177–1188, 2018. muscle performance improvement in both healthy humans [41] E. Barreiro, “Models of disuse muscle atrophy: therapeutic and athletes?,” European Journal of Applied Physiology, implications in critically ill patients,” Annals of Translational vol. 111, no. 10, pp. 2473–2487, 2011. Medicine, vol. 6, no. 2, p. 29, 2018. [24] S. Grosprˆetre, P. Gimenez, L. Mourot, and G. Coratella, [42] J. Kilby, K. Prasad, and G. Mawston, “Multi-channel surface “Elastic band exercise induces greater neuromuscular fatigue electromyography electrodes: a review,” IEEE Sensors Journal, than phasic isometric contractions,” Journal of Electromyog- vol. 16, no. 14, pp. 5510–5519, 2016. raphy and Kinesiology, vol. 47, pp. 113–120, 2019. [43] D. Devaprakash, G. J. Weir, J. J. Dunne, J. A. Alderson, and [25] S. C. F. A. Von Werder, T. Kleiber, and C. Disselhorst-Klug, C. J. Donnelly, “,e inﬂuence of digital ﬁlter type, amplitude “A method for a categorized and probabilistic analysis of the normalisation method, and co-contraction algorithm on surface electromyogram in dynamic contractions,” Frontiers clinically relevant surface electromyography data during in Physiology, vol. 6, no. 30, pp. 1–8, 2015. clinical movement assessments,” Journal of Electromyography [26] U. Imtiaz, L. Bartolomeo, Z. Lin et al., “Design of a wireless and Kinesiology, vol. 31, pp. 126–135, 2016. miniature low cost EMG sensor using gold plated dry elec- [44] Texas instrument, May 2018, http://www.ti.com/lit/ds/symlink/ trodes for biomechanics research,” in Proceedings of the 2013 ina128.pdf. IEEE International Conference on Mechatronics and Auto- [45] Texas instrument, June 2018, http://www.ti.com/lit/ds/symlink/ mation, pp. 957–962, Takamatsu, Japan, August 2013. tl072.pdf. [27] G. R. Jones, N. A. Neubauer, B. O’Connor, and J. M. Jakobi, ´ ´ ´ ´ [46] N. Miljkovic, N. Popovic, O. Djordjevic, L. Konstantinovic, “EMG Functional tasks recordings determines frailty phe- and T. B. Sekara, “ECG artifact cancellation in surface EMG notypes in males and females,” Experimental Gerontology, signals by fractional order calculus application,” Computer vol. 77, pp. 12–18, 2016. [28] S. Sadeghi Esfahlani, B. Muresan, A. Sanaei, and G. Wilson, Methods and Programs in Biomedicine, vol. 140, pp. 259–264, “Validity of the Kinect and Myo armband in a serious game 2017. for assessing upper limb movement,” Entertainment Com- [47] Analog devices, September 2018, https://www.analog.com/en/ puting, vol. 27, pp. 150–156, 2018. analog-dialogue/articles/practical-ﬁlter-design-precision-adcs.html. Journal of Healthcare Engineering 15 [48] A. Widmann, E. Schroger, ¨ and B. Maess, “Digital ﬁlter design voluntary contractions of human adductor pollicis,” Journal of Biomechanics, vol. 40, no. 1, pp. 1–8, 2007. for electrophysiological data - a practical approach,” Journal of Neuroscience Methods, vol. 250, pp. 34–46, 2015. [64] D. Roman-Liu and P. Bartuzi, “Inﬂuence of type of MVC test on electromyography measures of biceps brachii and triceps [49] A. F. T. Ibrahim, V. R. Gannapathy, L. W. Chong, and I. S. M. Isa, “Analysis of electromyography (EMG) signal for brachii,” International Journal of Occupational Safety and Ergonomics, vol. 24, no. 2, pp. 200–206, 2018. human arm muscle: a review,” in Lecture Notes in Electrical [65] J. J. D. Van der Gon, B. M. ter Haar Romeny, and Engineering, pp. 567–575, 2016. E. J. van Zuylen, “Behaviour of motor units of human arm [50] M. B. I. Raez, M. S. Hussain, and F. Mohd-Yasin, “Techniques muscles: diﬀerences between slow isometric contraction and of EMG signal analysis: detection, processing, classiﬁcation relaxation,” De Journal of Physiology, vol. 359, no. 1, and applications,” Biological Procedures Online, vol. 8, p. 11, pp. 107–118, 1985. [66] R. U. Newton, W. J. Kraemer, K. Hakkinen, ¨ B. J. Humphries, [51] SENIAM, March 2017, http://www.seniam.org/. and A. J. Murphy, “Kinematics, kinetics, and muscle acti- [52] F. Negro, K. Keenan, and D. Farina, “Power spectrum of the vation during explosive upper body movements,” Journal of rectiﬁed EMG: when and why is rectiﬁcation beneﬁcial for Applied Biomechanics, vol. 12, no. 1, pp. 31–43, 1996. identifying neural connectivity?,” Journal of Neural Engi- [67] S. Maeo, Y. Yoshitake, Y. Takai, T. Fukunaga, and neering, vol. 12, no. 3, article 036008, 2015. H. Kanehisa, “Neuromuscular adaptations following 12-week [53] K. Praveen, A. D. Arun, and M. Sivaramakrishna, “EMI/EMC maximal voluntary co-contraction training,” European qualiﬁcation of diﬀerential pressure monitoring system for Journal of Applied Physiology, vol. 114, no. 4, pp. 663–673, PFBR—A case study,” in Proceedings of the 2016 International Conference on ElectroMagnetic Interference & Compatibility (INCEMIC), pp. 1–4, Bangalore, India, December 2016. [54] J. Zhang, L. Wang, L. Yu, Y. Yang, Y. Zhang, and B. Li, “A low- oﬀset analogue front-end IC for multi-channel physiological signal acquisition,” in Proceedings of the 2009 Annual In- ternational Conference of the IEEE Engineering in Medicine and Biology Society, pp. 4473–4476, Minneapolis, MN, USA, September 2009. [55] C. K. Chui and G. Chen, “Kalman ﬁlter: an elementary ap- proach,” in Kalman Filtering, Springer, Cham, Switzerland, [56] A. Toprak and I. Guler, ¨ “Angiograph image restoration with the use of rule base fuzzy 2D Kalman ﬁlter,” Expert Systems with Applications, vol. 35, no. 4, pp. 1752–1761, 2008. [57] M. Schimmack, S. Nguyen, and P. Mercorelli, “Anatomy of haar wavelet ﬁlter and its implementation for signal pro- cessing,” IFAC-PapersOnLine, vol. 49, no. 6, pp. 99–104, 2016. [58] S. Roldan-Vasco, E. Perez-Giraldo, and A. Orozco-Duque, “Continuous wavelet transform for muscle activity detection in surface EMG signals during swallowing,” in Applied Computer Sciences in Engineering. WEA 2018. Communica- tions in Computer and Information Science, J. Figueroa- Garc´ıa, J. Villegas, J. Orozco-Arroyave, and P. Maya Duque, Eds., vol. 916, Springer, Cham, Switzerland, 2018. [59] S. Ranaldi, C. De Marchis, and S. Conforto, “An automatic, adaptive, information-based algorithm for the extraction of the sEMG envelope,” Journal of Electromyography and Ki- nesiology, vol. 42, pp. 1–9, 2018. [60] S. Mazumdar, A. Saikia, N. Sahai, S. Paul, and D. Bhatia, “Determination of signiﬁcant muscle in movement of upper limb using maximum voluntary contraction of EMG signal,” in Proceedings of the 2017 4th International Conference on Signal Processing and Integrated Networks (SPIN), pp. 96–99, Noida, India, February 2017. [61] B. N. Cahyadi, I. Zunaidi, S. A. Bakar et al., “Upper limb muscle strength analysis for movement sequence based on maximum voluntary contraction using EMG signal,” in Proceedings of the 2018 International Conference on Com- putational Approach in Smart Systems Design and Applica- tions (ICASSDA), pp. 1–5, Kuching, Malaysia, August 2018. [62] J. Paquin and G. A. Power, “History dependence of the EMG- torque relationship,” Journal of Electromyography and Kine- siology, vol. 41, pp. 109–115, 2018. [63] E. N. Rousanoglou, A. E. Oskouei, and W. Herzog, “Force depression following muscle shortening in sub-maximal International Journal of Advances in Rotating Machinery Multimedia Journal of The Scientific Journal of Engineering World Journal Sensors Hindawi Hindawi Publishing Corporation Hindawi Hindawi Hindawi Hindawi www.hindawi.com Volume 2018 http://www www.hindawi.com .hindawi.com V Volume 2018 olume 2013 www.hindawi.com Volume 2018 www.hindawi.com Volume 2018 www.hindawi.com Volume 2018 Journal of Control Science and Engineering Advances in Civil Engineering Hindawi Hindawi www.hindawi.com Volume 2018 www.hindawi.com Volume 2018 Submit your manuscripts at www.hindawi.com Journal of Journal of Electrical and Computer Robotics Engineering Hindawi Hindawi www.hindawi.com Volume 2018 www.hindawi.com Volume 2018 VLSI Design Advances in OptoElectronics International Journal of Modelling & Aerospace International Journal of Simulation Navigation and in Engineering Engineering Observation Hindawi Hindawi Hindawi Hindawi Volume 2018 Volume 2018 Hindawi www.hindawi.com Volume 2018 www.hindawi.com Volume 2018 www.hindawi.com www.hindawi.com www.hindawi.com Volume 2018 International Journal of Active and Passive International Journal of Antennas and Advances in Chemical Engineering Propagation Electronic Components Shock and Vibration Acoustics and Vibration Hindawi Hindawi Hindawi Hindawi Hindawi www.hindawi.com Volume 2018 www.hindawi.com Volume 2018 www.hindawi.com Volume 2018 www.hindawi.com Volume 2018 www.hindawi.com Volume 2018 http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Journal of Healthcare Engineering Hindawi Publishing Corporation http://www.deepdyve.com/lp/hindawi-publishing-corporation/design-and-validation-of-multichannel-wireless-wearable-semg-system-w0UkRJ00bN

Loading next page...

References (71)

H. Silva (2018)
Physiological Sensing Now Open to the World: New Resources Are Allowing Us to Learn, Experiment, and Create Imaginative Solutions for Biomedical Applications
IEEE Pulse, 9
(2019)
Muscle Monitor, Kuopio, FINLAND
J. Gon, B. Romeny, E. Zuylen (1985)
Behaviour of motor units of human arm muscles: differences between slow isometric contraction and relaxation.
The Journal of Physiology, 359
Geng Yang, Jia Deng, Gaoyang Pang, Hao Zhang, Jiayi Li, Bin Deng, Zhibo Pang, Juan Xu, Mingzhe Jiang, P. Liljeberg, Haibo Xie, Huayong Yang (2018)
An IoT-Enabled Stroke Rehabilitation System Based on Smart Wearable Armband and Machine Learning
IEEE Journal of Translational Engineering in Health and Medicine, 6
C. Chui, Guanrong Chen (1991)
Kalman Filter: An Elementary Approach
(2019)
MyoTrac3
A. Cisnal, Rubén Alonso, J. Pérez-Turiel, J. Fraile, V. Lobo, V. Moreno (2018)
EMG Based Bio-Cooperative Direct Force Control of an Exoskeleton for Hand Rehabilitation: A Preliminary Study
Converging Clinical and Engineering Research on Neurorehabilitation III
Simone Ranaldi, C. Marchis, S. Conforto (2018)
An automatic, adaptive, information-based algorithm for the extraction of the sEMG envelope.
Journal of electromyography and kinesiology : official journal of the International Society of Electrophysiological Kinesiology, 42
Yuehong Yin, Yan Zeng, Xing Chen, Yuan-zhi Fan (2016)
The internet of things in healthcare: An overview
J. Ind. Inf. Integr., 1
E. Zemková, D. Hamar (2018)
Sport-Specific Assessment of the Effectiveness of Neuromuscular Training in Young Athletes
Frontiers in Physiology, 9
G. Jones, N. Neubauer, Brian O'Connor, J. Jakobi (2016)
EMG Functional tasks recordings determines frailty phenotypes in males and females
Experimental Gerontology, 77
Manuel Schimmack, S. Nguyen, Paolo Mercorelli (2016)
Anatomy of Haar Wavelet Filter and Its Implementation for Signal Processing
IFAC-PapersOnLine, 49
E. Barreiro (2018)
Models of disuse muscle atrophy: therapeutic implications in critically ill patients.
Annals of translational medicine, 6 2
(2018)
Analog devices
(2017)
EMG 151015
(2019)
myontec
Jinyong Zhang, Lei Wang, Li Yu, Yabei Yang, Yuanting Zhang, Bin Li (2009)
A low-offset analogue front-end IC for multi-channel physiological signal acquisition
2009 Annual International Conference of the IEEE Engineering in Medicine and Biology Society
James Paquin, G. Power (2018)
History dependence of the EMG-torque relationship.
Journal of electromyography and kinesiology : official journal of the International Society of Electrophysiological Kinesiology, 41
A. Widmann, E. Schröger, B. Maess (2015)
Digital filter design for electrophysiological data – a practical approach
Journal of Neuroscience Methods, 250
(2019)
Mpower
B. Cahyadi, I. Zunaidi, S. Bakar, W. Khairunizam, S. Majid, Z. Razlan, M. Muhammad, M. Rudzuan, W. Mustafa (2018)
Upper Limb Muscle Strength Analysis For Movement Sequence Based on Maximum Voluntary Contraction Using EMG Signal
2018 International Conference on Computational Approach in Smart Systems Design and Applications (ICASSDA)
D. Roman-Liu, P. Bartuzi (2018)
Influence of type of MVC test on electromyography measures of biceps brachii and triceps brachii
International Journal of Occupational Safety and Ergonomics, 24
D. Devaprakash, G. Weir, J. Dunne, J. Alderson, C. Donnelly (2016)
The influence of digital filter type, amplitude normalisation method, and co-contraction algorithm on clinically relevant surface electromyography data during clinical movement assessments.
Journal of electromyography and kinesiology : official journal of the International Society of Electrophysiological Kinesiology, 31
Ric Lovell, Michael Knox, Matthew Weston, J. Siegler, Scott Brennan, Paul Marshall (2014)
Hamstring injury prevention in soccer: Before or after training?
Scandinavian Journal of Medicine & Science in Sports, 28
S. Maeo, Y. Yoshitake, Y. Takai, T. Fukunaga, H. Kanehisa (2014)
Neuromuscular adaptations following 12-week maximal voluntary co-contraction training
European Journal of Applied Physiology, 114
K. Biron, K. Englehart, P. Parker (2012)
A Study of Kalman Filtering Applied to Myoelectric Signal State Tracking
, 35
Troy Paddock (2019)
Germany
Contesting the Origins of the First World War
(2017)
Trigno lab
S. Saponara (2017)
Wearable Biometric Performance Measurement System for Combat Sports
IEEE Transactions on Instrumentation and Measurement, 66
S. Heywood, Y. Pua, J. McClelland, P. Geigle, A. Rahmann, Kelly Bower, R. Clark (2018)
Low-cost electromyography - Validation against a commercial system using both manual and automated activation timing thresholds.
Journal of electromyography and kinesiology : official journal of the International Society of Electrophysiological Kinesiology, 42
U. Imtiaz, L. Bartolomeo, Zhuohua Lin, S. Sessa, H. Ishii, K. Saito, M. Zecca, A. Takanishi (2013)
Design of a wireless miniature low cost EMG sensor using gold plated dry electrodes for biomechanics research
2013 IEEE International Conference on Mechatronics and Automation
Róisín Howard, R. Conway, A. Harrison (2018)
Muscle activity in sprinting: a review
Sports Biomechanics, 17
S. Grosprêtre, P. Gimenez, L. Mourot, G. Coratella (2019)
Elastic band exercise induces greater neuromuscular fatigue than phasic isometric contractions.
Journal of electromyography and kinesiology : official journal of the International Society of Electrophysiological Kinesiology, 47
(2018)
Interpreting Signal Amplitudes in Surface Electromyography Studies in Sport and Rehabilitation Sciences
Thiago Boaventura, Lisa Hammer, J. Buchli (2016)
Interaction Force Estimation for Transparency Control on Wearable Robots Using a Kalman Filter
iv) The data obtained have to be at a close accuracy to biomedical systems
F. Negro, K. Keenan, D. Farina (2015)
Power spectrum of the rectified EMG: when and why is rectification beneficial for identifying neural connectivity?
Journal of Neural Engineering, 12
Tyler Williams, Danilo Tolusso, Michael Fedewa, M. Esco (2017)
Comparison of Periodized and Non-Periodized Resistance Training on Maximal Strength: A Meta-Analysis
Sports Medicine, 47
M. Kos, I. Kramberger (2017)
A Wearable Device and System for Movement and Biometric Data Acquisition for Sports Applications
IEEE Access, 5
(2019)
MyoTrac, Montreal, Quebec, Canada
ii) Be able to ﬁlter the noise of movement during isotonic exercises and noise and distortions in SEMG signals appearing as a result of other factors
Sushmi Mazumdar, Angana Saikia, Nitin Sahai, S. Paul, D. Bhatia (2017)
Determination of significant muscle in movement of upper limb using maximum voluntary contraction of EMG signal
2017 4th International Conference on Signal Processing and Integrated Networks (SPIN)
R. Newton, W. Kraemer, K. Häkkinen, B. Humphries, A. Murphy (1996)
Kinematics, Kinetics, and Muscle Activation during Explosive Upper Body Movements
Journal of Applied Biomechanics, 12
Osman Özkaraca, A. Işık, I. Güler (2011)
Biological Signals Sensing with Wearable Technologies and an ECG Athlete Design
, 4
(2018)
Texas instrument
J. Kilby, K. Prasad, G. Mawston (2016)
Multi-Channel Surface Electromyography Electrodes: A Review
IEEE Sensors Journal, 16
A. Ranavolo, G. Chini, A. Silvetti, S. Mari, M. Serrao, F. Draicchio (2018)
Myoelectric manifestation of muscle fatigue in repetitive work detected by means of miniaturized sEMG sensors
International Journal of Occupational Safety and Ergonomics, 24
Shabnam Esfahlani, Bogdan Muresan, Alireza Sanaei, G. Wilson (2018)
Validity of the Kinect and Myo armband in a serious game for assessing upper limb movement
Entertain. Comput., 27
Adrian Burns, B. Greene, M. Mcgrath, Terrance O'Shea, B. Kuris, Steven Ayer, Florin Stroiescu, V. Cionca (2010)
SHIMMER™ – A Wireless Sensor Platform for Noninvasive Biomedical Research
IEEE Sensors Journal, 10
M. Vromans, P. Faghri (2017)
Electrical Stimulation Frequency and Skeletal Muscle Characteristics: Effects on Force and Fatigue
European Journal of Translational Myology, 27
Ryan Li, Scott Kling, Michael Salata, Sean Cupp, Joe Sheehan, J. Voos (2016)
Wearable Performance Devices in Sports Medicine
Sports Health, 8
I. Dhindsa, R. Agarwal, H. Ryait (2019)
Performance evaluation of various classifiers for predicting knee angle from electromyography signals
Expert Systems, 36
B. Schoenfeld (2010)
The mechanisms of muscle hypertrophy and their application to resistance training.
Journal of strength and conditioning research, 24 10
S. Roldán-Vasco, E. Pérez-Giraldo, A. Orozco-Duque (2018)
Continuous Wavelet Transform for Muscle Activity Detection in Surface EMG Signals During Swallowing
J. Gondin, P. Cozzone, D. Bendahan (2011)
Is high-frequency neuromuscular electrical stimulation a suitable tool for muscle performance improvement in both healthy humans and athletes?
European Journal of Applied Physiology, 111
Abdullah Toprak, I. Güler (2008)
Angiograph image restoration with the use of rule base fuzzy 2D Kalman filter
Expert Syst. Appl., 35
D. Clark, M. Lambert, A. Hunter (2018)
Contemporary perspectives of core stability training for dynamic athletic performance: a survey of athletes, coaches, sports science and sports medicine practitioners
Sports Medicine - Open, 4
iii) Its procedures like calibration, etc., have to continue for a short time
K. Praveen, A. Arun, M. Sivaramakrishna (2016)
EMI/EMC qualification of differential pressure monitoring system for PFBR — A case study
2016 International Conference on ElectroMagnetic Interference & Compatibility (INCEMIC)
E. Rousanoglou, Ali Oskouei, W. Herzog (2007)
Force depression following muscle shortening in sub-maximal voluntary contractions of human adductor pollicis.
Journal of biomechanics, 40 1
V. Bonaiuto, Nunzio Lanotte, C. Romagnoli, N. Silvaggi, G. Annino (2018)
Design of a Wireless Wearable DAQ System for the Evaluation of Sports Performances
, 2
(2019)
Viking on nicolet
B. Wallace, H. Bergstrom, T. Butterfield (2018)
Muscular bases and mechanisms of variable resistance training efficacy
International Journal of Sports Science & Coaching, 13
Jawad Hussain, K. Sundaraj, Y. Low, C. Lam, Sebastian Sundaraj, Md. Ali (2018)
A systematic review on fatigue analysis in triceps brachii using surface electromyography
Biomed. Signal Process. Control., 40
A. Ibrahim, V. Gannapathy, L. Chong, Ida Isa (2016)
Analysis of Electromyography (EMG) Signal for Human Arm Muscle: A Review
S. Werder, T. Kleiber, C. Disselhorst-Klug (2015)
A method for a categorized and probabilistic analysis of the surface electromyogram in dynamic contractions
Frontiers in Physiology, 6
M. Reaz, M. Hussain, F. Mohd-Yasin (2006)
Techniques of EMG signal analysis: detection, processing, classification and applications
Biological Procedures Online, 8
M. Cardinale, M. Varley (2017)
Wearable Training-Monitoring Technology: Applications, Challenges, and Opportunities.
International journal of sports physiology and performance, 12 Suppl 2
N. Miljković, Nenad Popovic, O. Djordjevic, L. Konstantinović, T. Šekara (2017)
ECG artifact cancellation in surface EMG signals by fractional order calculus application
Computer methods and programs in biomedicine, 140
R. Panariello, Timothy Stump, F. Cordasco (2017)
The Lower Extremity Athlete: Postrehabilitation Performance and Injury Prevention Training
Operative Techniques in Sports Medicine, 25
Sebastian Skals, J. Vinstrup, E. Sundstrup, M. Jakobsen, C. Andersen, L. Andersen (2018)
Shoulder and arm muscle activity during elastic band exercises performed in a hospital bed
The Physician and Sportsmedicine, 46

Publisher: Hindawi Publishing Corporation
Copyright: Copyright © 2019 Serkan Örücü and Murat Selek. 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: 2040-2295
eISSN: 2040-2309
DOI: 10.1155/2019/4580645
Publisher site: See Article on Publisher Site

Abstract

Hindawi Journal of Healthcare Engineering Volume 2019, Article ID 4580645, 15 pages https://doi.org/10.1155/2019/4580645 Research Article Design and Validation of Multichannel Wireless Wearable SEMG System for Real-Time Training Performance Monitoring 1 2 Serkan Oru¨cu¨ and Murat Selek Ermenek Vocational School, Karamanog˘lu Mehmetbey University, Karaman 70400, Turkey Vocational School of Technical Sciences, Konya Technical University, Konya 42130, Turkey Correspondence should be addressed to Serkan Oru¨cu¨; srknorucu@kmu.edu.tr Received 8 April 2019; Revised 20 July 2019; Accepted 16 August 2019; Published 9 September 2019 Guest Editor: Federica Verdini Copyright © 2019 Serkan Oru¨cu¨ and Murat Selek. ,is 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. Monitoring of training performance and physical activity has become indispensable these days for athletes. Wireless technologies have started to be widely used in the monitoring of muscle activation, in the sport performance of athletes, and in the examination of training eﬃciency. ,e monitorability of performance simultaneously in the process of training is especially a necessity for athletes at the beginner level to carry out healthy training in sports like weightlifting and bodybuilding. For this purpose, a new system consisting of 4 channel wireless wearable SEMG circuit and analysis software has been proposed to detect dynamic muscle contractions and to be used in real-time training performance monitoring and analysis. ,e analysis software, the Haar wavelet ﬁlter with threshold cutting, can provide performance analysis by using the methods of moving RMS and %MVC. ,e validity of the data obtained from the system was investigated and compared with a biomedical system. In this comparison, 90.95%± 3.35 for left biceps brachii (BB) and 90.75%± 3.75 for right BB were obtained. ,e output of the power and %MVC analysis of the system was tested during the training of the participants at the gym, and the training eﬃciency was measured as 96.87%± 2.74. contraction and relaxation of muscles. When academic 1. Introduction studies related to this subject are analysed, there are some In recent years, the monitoring of athlete performance has wearable biometric systems developed for the purpose of become indispensable for the health of athletes. Wireless the monitoring of performance during training. Some of technologies have started to be widely used in order to obtain these systems are intended for recording parameters like data for the purpose of examining training eﬃciency in the heart rate, respiration, location, and velocity or for esti- monitoring of muscle activation and sport performance of mating the levels of muscle fatigue [10–13]. Some of them athletes [1, 2]. It is possible to collect information about have been produced for the measurement of the SEMG athlete performance and rehabilitation, about preventing signals in laboratorial environment [14]. Another pro- muscle fatigue or injuries through posttraining analysis of portion of them has been designed for the purpose of perceiving dynamic muscle contraction during isolated SEMG signal obtained during the training [3–5]. Recording of SEMG signals in related muscles during training can be training through the SEMG [15]. ,e last proportion has extremely useful in increasing performance and preventing carried out low-cost experimental SEMG systems and disabilities [6]. matched the key features of the system with the existing Traits of SEMG signals obtained during training (fre- systems [16–18]. quency, severity, etc.) change depending on the muscle ,e most reliable method used in the adequacy and group measured and the severity of contraction [7–9]. In examination of muscle activation in physiological studies is these measurements, surface-type electrodes are used to the amplitude analysis carried out on SEMG signals, known determine and examine the activity of muscles during as MVC (maximum voluntary contraction) normalization 2 Journal of Healthcare Engineering like weightlifting and bodybuilding, for the performance [19]. Data with MVC normalization enable understanding of what capacity the muscle works, how eﬀective level evaluation of the athlete until the motor skills of the movement are improved and at necessary moments in muscles have reached through training and how much eﬀort a training requires from an athlete [20]. preventing the injury process by intervening in training. ,e simultaneous monitorability of athlete performance Based on these elements, a new wireless wearable SEMG during the process of training is a must for athletes at the data collection system has been introduced which enables beginner level to being able to carry out healthy training in performance monitoring and analysis at training time with sports like weightlifting and bodybuilding [21, 22]. ,is its real-time MVC normalization and contraction detection feature enables performance evaluation to be carried out processes. ,e SEMG circuit used in our system is designed momentarily during the time when there is no trainer or until by us to be used in future studies and to be developed the motor skills of the athlete concerning movement develop according to our needs. In the presented system, digital ﬁltering is also used in enough. A SEMG system, to be used during the training for this purpose [7, 23–25] has to addition to hardware ﬁltering in SEMG circuit. ,ese nu- merical ﬁlters are Haar wavelet ﬁlters with ,reshold cutting (i) Be able to provide the required SEMG data necessary based on (TCHW) and linear Kalman [37, 38]. Each nu- for monitoring training eﬃciency in performance merical ﬁltering method is tested together with hardware analysis ﬁltering. Results obtained from here will be determinative in (ii) Be able to ﬁlter the noise of movement during deciding the ﬁltering structure that can be used in future isotonic exercises and noise and distortions in stages of the system design. Subsequently, ﬁltered data are SEMG signals appearing as a result of other factors processed through moving RMS method containing the methods of moving average (MA) and root mean square (iii) Its procedures like calibration, etc., have to continue for a short time (RMS), scaled through MVC normalization, and a training support system that can carry out real-time performance (iv) ,e data obtained have to be at a close accuracy to analysis and monitoring. biomedical systems (v) Has to be simultaneously usable in a training 2. Materials and Methods environment For use in the industrial ﬁeld, various systems are 2.1. Isotonic Contraction. Isotonic contraction encompasses available for SEMG data collection and processing. To exercises where muscle tendons get shortened to generate investigate these, WB-EMG [26], BiometricsDatalog [27], movement. Any kind of movement, ranging from weight- Myo Armband [28], DelsysTrignio [29], BITalino [30], lifting to rowing and running, is in this category [39]. In Mbody3 [31], Mpower [32], MyoTrac [33], MyoWare [34], sport, an isotonic exercise is a training where the most Shimmer [35], and hospital [36] are such systems. ,e amount of strength is exerted on a particular muscle or systems speciﬁed in [26], [27], and [29] and the systems muscle group to increase that muscle mass or performance which we measure in hospital [36] are not wearable during in general. Due to the fact that human activity and athletic training. ,e system speciﬁed in [44] is wearable and performance necessitate these kinds of movement, isotonic supports wireless transmission but its production is exercises form the basis of a lot of training protocols [40]. It stopped. In terms of the electrodes used, and CMRR, there is possible to observe pathological changes or eﬃciency is no diﬀerence in all of these products and they comply obtained from the training through an examination of with the SENIAM criteria. ,e systems [26], [27], and [30] SEMG signals generated in muscles during these exercises do not have noise and data processing ﬁlters, and the [41]. systems in [28] and [29] use a Notch ﬁlter and a band-stop ﬁlter with narrow-bandwidth in hardware. ,e system in [26, 34] is designed for single-channel use but does not 2.2. SEMG Circuit Design. ,e SEMG circuit design details support multichannel monitoring. ,e systems in [31–34] are given below. ,e circuit consisting of 4 channels could are wearable and do not include contraction detection and monitor the biopotential change of 4 diﬀerent muscle groups simultaneous MVC analysis although they can monitor at the same time. So, it is possible to monitor biopotential multiple muscle groups. A summary of these comparisons changes occurring in muscles in symmetrical movements is presented in Table 1. that aﬀect multiple muscle groups (e.g., the Bench Press When the table is analysed, it is seen that all of these movement aﬀects pectoralis major and triceps muscles). ,e systems can simultaneously observe biopotential changes in circuit has in each channel, respectively, one in- muscle or muscle groups monitored during training, but none strumentation ampliﬁer, a inverting ampliﬁer, a low-pass of them include real-time MVC normalization and con- ﬁlter, a high-pass ﬁlter, and a full-wave rectiﬁer. ,e circuit traction detection procedures for performance analysis has a diode for input protection, a pointer indicating that the during training. circuit is working, and a start-up button. During working, ,at these features can be monitored simultaneously the LD1117 regulator was used for the Bluetooth feed and during the training process may be useful especially for the 7805 regulator for the +5 volt and − 5 volt op-amp feed beginner athletes to perform a healthy training in sports (Figure 1(a)). ,e SEMG signals we want to process are Journal of Healthcare Engineering 3 Table 1: Comparison of the SEMG acquisition systems. Real- Number ADC Signal Contraction time Connection System of Gain resolution Wearable Filter type CMRR type detection MVC type channels (bits) norm. Proposed SEMG 4 4400 12 Yes Hardware + software Yes Yes >90 Bluetooth system WB-EMG SEMG 1 100–10000 12 No No No No >90 Bluetooth Biometrics SEMG 8 1000 14 No No No No >90 Bluetooth datalog Myo SEMG 8 ≥1000 8 Yes Notch No No >90 Bluetooth armband Delsys SEMG 16 909 16 No Notch No No >90 RF Trignio BITalino SEMG Up to 6 1000 6–10 Yes No No No >90 Bluetooth Mbody3 SEMG Up to 6 ≥1000 24 Yes Hardware + software No No >90 Bluetooth Mpower SEMG 4 ≥1000 — Yes Hardware + software No No >90 Bluetooth MyoTrac SEMG 2 ≥1000 14 Yes Butterworth No No >90 Bluetooth MyoWare SEMG 1 ≥1000 — Yes No No No >90 Bluetooth Shimmer SEMG Up to 60 ≥1000 16 Yes Hardware + software No No >90 Bluetooth Hospital SEMG 8 1–10000 24 No Hardware + software Yes No >90 Usb MUAP signals whose amplitude is between 0 and 1.5 mVolt changed depending on the input signal parameters [48]. In (RMS). To process this electrical signal, it must ﬁrstly be these applications, analogue ﬁlters are used to eliminate noise from the signal in signal ampliﬁcation and processing ampliﬁed. In the system, this ampliﬁcation is done by in- creasing the diﬀerence between the two electrodes in bipolar circuits, to provide noise immunity, and to obtain the mode. While the obtained common signal is ampliﬁed in this necessary parts of the frequency band [49]. On the contrary, mode, the background noise is also suppressed. Two of the digital ﬁlters are used to ﬁlter signal residues named artifact probes activated from each channel are connected to the after motion and to analyse SEMG signal (feature extraction, circuit’s soil, like the reference probe [42] which is placed in time-frequency analysis, contraction detection, performance a more electrically remote area (preferably a neutral or close analysis, etc.) [41, 50]. to the bone region) while going to the ampliﬁer and ﬁlter In the circuit, analogue ﬁltering is performed by low- and circuits over INA 128P, which operates in a single diﬀer- high-pass ﬁlters. Ideal SEMG signals are observed between 50 Hz and 500 Hz and should be ﬁltered from frequency ential mode. In the ﬁrst step, ampliﬁcation was performed by using the INA 128P diﬀerential ampliﬁer (Figure 1(b)). components outside this range [51]. For this, the signal from the output of the instrumentation ampliﬁer is ﬁrst ﬁltered so As stated in [43], the reason why we use INA 128P is that the amplitude of the SEMG signal is low and that the that the gain is 1 in the high-pass ﬁlter (HPF) using TL072 ampliﬁer to be used due to other factors like noise must have with a cutoﬀ frequency of about 48 Hz (Figure 1(d)). ,e a high input impedance and a high common mode rejection components of the EMG signal above 500 Hz are ﬁltered rate (CMRR> 95 dB). ,is ampliﬁer has the required fea- through a 2nd order Sallen–Key low-pass ﬁlter (LPF) using tures with CMRR >120 dB and 10 GΩ input impedance. TL072. ,rough this section, resistance and capacitor values When we set the gain value for the 60 Hz input signal to are designed so that the cutoﬀ frequency is approximately G � 74.52 using INA 128P in our system, approximately 482 Hz, the quality factor is 0.5, and the gain is 1 108 dB CMRR was obtained as stated in the technical (Figure 1(e)). ,e reason we prefer the Sallen–Key topology we use in the circuit is that this ﬁlter has the ability to document in [44]. ,e reason for selecting a 60 Hz input signal in the system design is that the SEMG signal is produce a quadratic low-pass reaction with better selectivity (higher pole) and various possible approaches (Butterworth, dominant in the range of 50 Hz to 150 Hz. To obtain a processable signal amplitude in the second stage, TL072 was Chebyshev, ,omson-Bessel, etc.) [43, 47, 49]. ,is will help used as shown in Figure 1(c) as an active inverting ampliﬁer. us in our future work. At this stage was the ampliﬁer gain approximately G � 59 ,en, the whole SEMG signal was moved to the positive and the CMRR approximately 100 dB by using the 60 Hz level using the full-wave rectiﬁer (Figure 1(f)). With this input signal as stated in [45]. process, it is possible to analyse the low-frequency oscilla- In SEMG applications, analogue (hardware) and digital tions by overcoming the high-pass nature of the SEMG (software) ﬁlters are used to remove unwanted component signal [52]. ,us, it is aimed to use the circuit except for the noises and process the necessary parts in the SEMG signal training eﬃciency, also in the ﬁelds of prosthesis control and ergonomics. [46]. Analogue ﬁlters remove anything above or below a selected cut frequency, while digital ﬁlters make this process ,e Pic16F1786 microcontroller with connected full-wave rectiﬁer outputs contains 11 12 bit A/D (Analogue/Digital) more precise as they can be programmed [47]. ,is certainty is due to the fact that the features of digital ﬁlters can be converters. ,e data obtained from the rectiﬁer of each channel 4 Journal of Healthcare Engineering U2 U7 7805 LD1117V33 D1 U3 1 3 3 2 StereoJack R1 VI VO IN OUT J1 1N4007 680 2 inaOut GND U1 C2 C4 C11 C8 C9 RG1 C10 1 6 7805 220uF 220uF 100uF 1 100uF RG2 J2 2 13 5 2 V1 100uF 100uF VO REF C3 INA122 C1 POWER 220uF 220uF –5V (a) (b) C6 R3 R5 330k 33k 1nF U4:B U5:A U4:A R7 R6 + + C5 3 7 1 AMPOUT R4 + HPFOUT 330k 330k 2 1 6 R2 – INAOUT – hpfOut LPFOUT AMPOUT 33k 100nf C7 5.6k 1nF TL072 TL072 TL072 (c) (d) (e) R8 U8 10k RECTOUT RA0 RC0 Tx 3 12 U6:A Rx U5:B RA1 RC1 4 13 D2 RA2 RC2 1N4148 3 RA3 RC3 +3.3V 6 15 + 1 Vcc RA4 RC4 GND 7 16 R10 2 GND 6 – HC06 RA5 RC5 – RectOut TX Tx 10 17 RA6 RC6 Rx RX 10k 9 18 LPFOUT RA7 RC7 D3 TL072 TL072 21 25 1N4148 RB0 RB4 R11 R9 R12 RB1 RB5 23 27 RB2 RB6/ICSPCLK 24 28 RB7/ICSPDAT RB3 10k 10k 10k RE3/MCLR/VPP PIC16F1786 (f ) (h) (g) fe d c b In. (i) Figure 1: Block diagram and mounted state of the SEMG circuit. (a) Regulator circuit. (b) Instrumentation ampliﬁer. (c) Inverting st nd ampliﬁer. (d) 1 -order HPF. (e) 2 -order Sallen–Key LPF. (f) Full-wave rectiﬁer. (g) PIC 16F1786. (h) Bluetooth module. (i) Mounted state of the SEMG circuit. in the circuit are connected, respectively, to the RA0-RA3 received and processed by the data collection program inputs of this controller. ,is microcontroller performs the written in the C# language. ,e digitalized SEMG data in A/D conversion in 20 ms time intervals through the pro- the data collection program are processed through digital gram we write. ,e converted channel data are turned into ﬁlters. ,e PCB (printed circuit board) of the circuit is a string, and this sends data from the RC0 output to the designed to be 10 cm × 10 cm in size, and as stated in [53], the Bluetooth module (Figure 1(g)). ,e transmitted data have PCB tracks are intended to be exposed to as little noise as a resolution of 2.4 μV in each step. Data sent at 4800 bps possible. ,e mounted state of the circuit shown in Figure 1(i) speed via the HC-06 Bluetooth module (Figure 1(h)) are is boxed and placed inside a wearable belt. ,e necessary –5V +5V 4 8 –5V 2 4 8 GND +5V GND +5V –5V 4 8 +5V 4 8 –5V +5V +3.3V 4 8 –5V +5V –5V +5V Journal of Healthcare Engineering 5 Table 2: Information about age, gender, weight, and height of the energy for the operation of the circuit was obtained from subjects. 1000 mAh lithium batteries. It is intended to minimize power line interference (PLI) without the need for any insulation, as Participant no. Age Gender Weight (kg) Height (cm) stated in [54] using battery in the system. 1 21 Male 80 163 2 25 Male 82.3 178 3 29 Male 87 180 2.3. Participants and Setup. Five males and two females 4 33 Male 85 177 voluntarily participated in our study and have at least 2 years 5 37 Male 104.6 193 of experience in strength training. ,e information of the 6 24 Female 70 180 participants is shown in Table 2. 7 27 Female 68 172 ,e participants were informed about the content of our study, and a signed consent form was obtained from all of them. All exercises and measurements were made under the supervision of a specialized trainer. As described in the recommendations of the European initiative known as SENIAM (surface electromyography for noninvasive muscle evaluation of muscles) by selecting 10 mm diameter electrodes shown in Figure 2 for SEMG, the bipolar con- ﬁguration is located 1–2 cm away from the centre of the muscle and the reference electrode is placed in a region that is electrically neutral according to the action [51]. ,e connection between the electrodes and the circuit channels is provided by using armoured cables which have 3.5 mm Figure 2: Example view of electrodes and shielded cables. ends, 3 colour code (red, green, and blue) and labelled contacts (L, F, and R), as shown in Figure 2. When the system is modelled, it was aimed to minimize the Our experiments consist of 3 parts. In the ﬁrst part, 8 distortions in data by estimating the k parameter speciﬁed by repetitions and 1 set of alternate dumbbell curl (ADBC) x in SEMG data array at a particular time as X : training was performed using a maximum load of 60–70%. 􏽢 􏽢 X � K · Z + (1− )K · X . (1) In this section, ﬁrstly, it is investigated whether the ana- k k k k k− 1 logue ﬁlter data obtained from the circuit in the training Here, Z expresses the measuring data wanted to be reﬂect the biopotential activity changes that occur during absolutized, K the Kalman gain and X the measuring k k− 1 the training. In the sequel, the analogue ﬁlter data obtained data belonging to the previous stage. If the system is from the circuit are processed by means of Kalman and modelled through this information, a model consisting of threshold cut Haar wavelet ﬁlter (TCHW) to eliminate calculation (2) and update (3) is obtained. noise sources and to investigate the perceptibility of the x � Ax + Bu + w , (2) isotonic contractions. k k− 1 k k− 1 In the second part, the accuracy of the developed system was compared with the biomedical system (Viking on z � Hx + v . (3) k k k Nicolet EDX) used in Karaman State Hospital (See Table 1). In (2), any x is expressed as a linear combination of the In this comparison, the RMS values obtained from both k next control signal k of its previous value and the noise of the systems were used. process. In (3), any measurement value making certain of the In the third part, the availability of moving RMS and accuracy of which we are not sure is accepted to be a linear %MVC values as the screen output of the system was combination of the signal value and the noise of the investigated in terms of performance feedback. For this measurement. purpose, ﬁrst, the moving RMS values obtained by asking In HW, the main wavelet acts as the wavelet transform users to perform a second ADBC (8 repetitions 1 set) but is scaled and shifted during this procedure of wavelet training were recorded. In addition, a %MCV mea- transform [35]. Scaling corresponds to the widening and surement was made by asking all users in the training constriction of the signal (f(t)) and the shift to the wave environment to lift 5 kg dumbbell and maximum weight shift (τ) in the timescale axis (t) in the following equation (Men 17.5 kg, 20 kg, and 25 kg dumbbell; women 12.5 kg [57, 58]: and 15 kg dumbbell) they can. − jωt F(ω, τ) � 􏽚 f(t)w(t − τ)e dt. (4) 2.4. Kalman and TCHW Filters. Kalman ﬁlter is used to estimate the system status from input and output in- formation with the previous information of a model in a HW is a wavelet-based, scaled, “square-shaped” array of dynamic system indicated by the state-space model [55, 56]. functions. ψ(t), the main function of HW (5), and also φ(t), 6 Journal of Healthcare Engineering 􏽳�� a scaling function (6), are deﬁned in t time interval given as (11) follows: f � [f(t)] dt, rms T − T 2 1 1 ⎧ ⎪ ⎪ 1, 0≤ t≤ , ⎪ 􏽳�� ⎪ 2 ⎨ (12) f � lim 􏽚 [f(t)] dt. rms T⟶∞ ψ(t) � (5) T 0 − 1, < t≤ 1, Another method we use as MA is the technique of analysing changes in a data set to estimate long-term trends. 0, otherwise, For a given N time window, if the values s , s , s ,. . ., s 1 2 3 n corresponding to this time interval of the S variable shown in ⎧ ⎪ 1, 0≤ t≤ , ⎨ the times t , t , t , . . ., t are known, the MA window size is 1 2 3 n φ(t) � (6) ⎪ deﬁned as N � 2k + 1 and processed as speciﬁed in 0, otherwise, +k MA � s . ,e Haar function ψ is deﬁned as shown in 􏽘 (13) n,k i− j j�− k n/2 n ψ (t) � 2 ψ 2 t − k , t ∈ R. (7) n,k ,us, changes in the time window given at the j moment are obtained by averaging the time series of the k Since the SEMG signals are user-based, SEMG signals between isotonic muscle contractions may vary according to time in the j moment. Instead of using the above- mentioned RMS and MA methods separately, the moving the individual. In the method we use with HW, the indi- vidual waits for approximately 2–4 seconds with the weight RMS method was used in our system by calculating the RMS value in a moving window, which is a combination of in his hand before starting training and in the meantime, the procedure of threshold cutting in the system can be carried these methods. In this method, the operation can be performed at any t time interval of the moving window; out. ,e threshold cutting is based on the calculation of the average value (8), the standard deviation (9), and the signal therefore, it acts as a ﬁlter in a certain time interval, as slope (10): shown in (14). In this way, the processing of the data obtained according to the variable speed of the replays in A � ∗ 􏽘 x , (8) i the training sets gets easier. In this equation, n refers to the i�1 length of the window, while x(k) refers to the data within 􏽶�� the window: 1/2 (9) σ � 􏽘 x − μ􏼁 , i ⎝ ⎠ ⎛ ⎞ i�1 (14) x [i] � 􏽘 x [k] . RMS j�(i− N+1) 􏽐(x − x)(y − y) So, it can be measured how much power is obtained from s � . (10) 􏽐(x − x) the muscle through the moving RMS value. ,e MVC (maximum voluntary contraction-maximum Here, x is the value added to the average, μ is the average amplitude of the signal) normalization is widely used in value and N is the number of the total value. After the values SEMG signals as an amplitude analysis technique. ,e re- of the average, standard deviation and slope are calculated sults are shown as a percentage (%MVC) of the MVC value and all SEMG signals complying with this condition are that can be used to create a common background when equalled to zero. ,us, the signals between the voluntary comparing data between subjects [60, 61]. SEMG signals contractions can be eliminated. depend on the user and have a structure that can cause records to change even when measured from the same 2.5. RMS, MA, and %MVC. After the SEMG signal is cap- position with the same motion. ,erefore, MVC normali- tured, the commonly used RMS or MA values are analysed zation is used to eliminate this diﬀerence and to enable data by using [59]. In RMS analysis, the SEMG signal is subjected comparison between subjects [61]. MVC expresses the to a set of mathematical operations designed to measure the highest value obtained in a repeat during this measurement power of change. ,us, the intensity and duration of events to normalize SEMG signals obtained for a speciﬁc purpose, like muscle contractions can be investigated. ,erefore, the while SMVC (submaximal voluntary contraction) refers to RMS value is a parameter chosen during contraction and the voluntarily recorded SEMG data. %MVC corresponds to reﬂects the level of physiological activity in the body. the multiplication of the normalized value of according to Mathematically, the RMS value of a continuous-time SMVC’s MVC with 100 [62, 63]: waveform is the square root of a function deﬁning the SMVC continuous waveform shown in f (t) in the following, de- %MVC � (15) 􏼒 􏼓∗ 100. ﬁned in the range T ≤ t≤ T : MVC 1 2 Journal of Healthcare Engineering 7 (a) (d) MVC analysis 4x Surface EMG inst. amp. Amplifier 1-order HPF EMG electrode G = 74.5 G = 59 M-RMS calculation Haar wavelet filter 2-order 12 bit A/D Bluetooth Sallen–Key LPF converter module PC (b) (c) Figure 3: Overview of the system. (a) Connecting electrodes before training (Photoshoot by Orucu). (b) Block diagram of the SEMG circuit. (c) Block diagram of the analysis software. (d) User interface of the analysis software. ,us, it can be scaled how much power is obtained from the training speed (Figure 3(c)). After this process, the the muscle or muscle groups investigated in repetitions in SMVC value of each repetition in each set of the training is processed according to the previously saved MVC values. each set of training. ,en, %MVC values are displayed on the screen in separate graphs according to the channels from which the data are 2.6. Proposed System. Our system has the ability to follow taken. Finally, they are saved to the database in “.csv,” the biopotential changes of four diﬀerent superﬁcial muscle “.dat,” and “.xlsx” formats (Figure 3(d)). groups at the same time. ,e reason why the system is designed with 4 channels is that most movements used in 3. Results and Discussion bodybuilding and weight training activate at least 1 or 3 muscle groups at the same time. ,e system takes the 3.1. Analogue + Digital Filtered Data from the System. ,e biopotential signals of the muscles that are activated during analogue-ﬁltered data of the ﬁrst 4 repetitions of ADBC training through surface electrodes (Figure 3(a)), and then, training performed by participant number two is shown in ﬁrst it ampliﬁes them in the instrumentation and ampliﬁer Figure 4(a), marked as 4(a) and 4(b) for each repetition. st parts in the SEMG circuit, after it ﬁlters them with the 1 - ,e left BB (LBB-Left Biceps Brachii) data are obtained nd degree high pass and 2 -degree Sallen–Key low-pass ﬁl- from CH1 (ﬁrst channel of the SEMG circuit), and the right ters. ,ese analogue-ﬁltered signals are sent to the com- BB (RBB-Right Biceps Brachii) data are obtained from CH2 puter via Bluetooth after a 12 bit analogue-to-digital (the second channel of the SEMG circuit). From the data conversion (Figure 3(b)). By the software we developed in obtained, some decrease in Rep2b, Rep3a, Rep3b, and Rep4a C# language, all SEMG channel data received by the (between 100 and 200 μV) and a data change during pushing computer are digitally ﬁltered and then they calculated the the weight down (relaxation period of the muscle) in Rep 4b moving RMS values in time windows that vary according to were observed. As we consulted with the professor of 8 Journal of Healthcare Engineering Rep2a Rep1b Rep4a Rep3b Rep1a Rep3a Rep2b Rep4b Time (millisecond) CH1-left biceps brachii CH2-right biceps brachii (a) Rep1 Rep2 Rep3 Rep4 Rep5 Rep6 Rep7 Rep8 Rep9 Rep10 Rep11 Rep12 Time (millisecond) CH3 CH4 (b) Figure 4: Sample analogue ﬁltered data obtained from the SEMG circuit during training: (a) Sample results of participant number two, (b) sample results of participant number six. ˘ both relax normally in Rep10, in which distortion in Physical Education and Sports Teaching (Karamanoglu Mehmetbey University), he stated that the fall was caused by movement appears as a result of fatigue in Rep11 and the distortion of movement. According to the consultant Rep12. In addition, the data of other participants obtained professor, this change appeared to have been caused by the from these trainings are presented in Figure 5. prolongation of the activation period of the muscle as a In Figure 6, the data, processed with TCHW and Kalman result of pushing the weight down more slowly as speciﬁed ﬁlters, of two repetitions in training, belonging to the right in [64, 65]. BB muscle, conducted by the participant numbered 4, are Other data of training performed by participant shown. In this Figure, 6(a) shows the analogue ﬁltered state number four are shown in Figure 4(b). In this training, LBB of the SEMG signal, and 6(b) shows the preliminary mea- data were obtained from CH3 (the third channel of the surement of the threshold cut-out. ,e average and standard SEMG circuit) and RBB data were obtained from CH4 (the deviation measured here were found as 61.11± 51.61 μV, and the slope was found as 0.005⁰. ,e signal ﬁltered with TCHW fourth channel of the SEMG circuit). When the results obtained are investigated in accordance with contraction after this procedure is shown in 6(c), and the signal pro- and relaxation situations as speciﬁed in [65, 66] which cessed through Kalman ﬁlter is shown in 6(d). Filtering consultant professor pointed, it is observed that BB muscles results indicate that the TCHW method produces better contract and relax normally in Rep1, Rep5, Rep7, and Rep9 results in ﬁltering unwanted signals and contraction de- and BB muscles contract fast and relax normally in Rep2. It tection compared to the method of Kalman ﬁlter. As a result is observed that the left BB contracts more than the right BB of these processes, it was decided to use TCHW ﬁlter in our does and both relax normally in Rep3, that the required system. support is taken from other regions and movement is ruined in Rep4 and that the left BB muscle contracts more, the right BB muscle contracts normally and both relax 3.2. Comparison Results with the Existing Biomedical System. ,e accuracy of the data obtained from our system was normally in Rep6 and Rep8. It is observed that the left BB compared through the data belonging to two men and two contracts normally and the right BB contracts more and SEMG signal (μVolt) SEMG signal (μVolt) 1009 Journal of Healthcare Engineering 9 Time (millisecond) CH1-left biceps CH3-right biceps (a) Time (millisecond) CH2-left biceps CH4-right biceps (b) Time (millisecond) CH1-right biceps CH4-left biceps (c) Figure 5: Continued. SEMG signal (μVolt) SEMG signal (μVolt) SEMG signal (μVolt) 1 1 9 9 17 17 33 36 37 41 56 53 61 65 65 69 81 77 81 91 109 117 113 121 117 125 136 129 129 146 133 151 161 153 10 Journal of Healthcare Engineering Time (millisecond) CH1-left biceps CH4-right biceps (d) Time (millisecond) CH2-left biceps CH3-right biceps (e) Figure 5: Data of other participants obtained from these trainings. (a) Results of participant number one. (b) Results of participant number three. (c) Results of participant number ﬁve. (d) Results of participant number six. (e) Results of participant number seven. 1200 250 CH3-left BB Raw data 0 0 0 2000 4000 6000 8000 0 10002000300040005000 t as millisecond t as millisecond (a) (b) Figure 6: Continued. Mean CH3-left BB as μV raw SEMG signal (μVolt) SEMG signal (μVolt) SEMG noise as uV 39 51 53 53 55 55 57 57 59 81 Journal of Healthcare Engineering 11 1200 525 CH3-left BB 400 110 0 2000 4000 6000 8000 t as millisecond t as millisecond (c) (d) Figure 6: Comparison of the ﬁltering results. (a) SEMG data without the ﬁlter. (b) Premeasurement for threshold ﬁlter. (c) SEMG signal with threshold + HW ﬁlter. (d) SEMG signal with Kalman ﬁlter. (a) (b) Figure 7: (a) A measurement taken in the hospital environment and a photograph of the current biomedical system. (b) A photograph taken at the gym before training. Table 3: Moving RMS Results in Gym and Hospital. Note that “M” denotes the measurement number; “BB” denotes biceps brachii; “S” denotes system; “H” denotes hospital, “MN” denotes muscle name. Participants/weight (no./kg) M MN Type 1/idle 2/idle 3/idle 4/idle 1/5 2/5 3/5 4/5 1/25 2/25 3/15 4/12.5 S 70.69 69.72 51.18 43.82 123.69 129.54 97.54 93.64 914.7 935.98 566.98 547.64 Left BB H 67.13 72.31 47.24 45.9 137.42 141.94 108.66 101.05 950.94 1112.53 616.53 604.36 S 71.4 69.75 49.66 42.45 119.11 127.41 96.86 93.95 960.71 937.69 565.69 515.43 Right BB H 69.64 70.51 50.22 43.93 135.57 143.13 107.93 97.14 943.82 1117.15 615.15 545.64 S 69.86 68.84 52.03 43.15 121.82 128.9 95.71 93.8 907.35 934.5 563.5 518.06 Left BB H 70.39 69.61 51.76 43.75 138.87 139.69 105.78 101.55 942.14 1116.89 614.89 595.59 II S 69.84 71.34 49.01 46.68 122.96 126.95 96.5 93.61 950.6 932.61 562.61 526.48 Right BB H 71.82 67.83 50.25 43.27 136.52 142.8 106.73 102.46 1002.4 1110.94 612.94 598.81 S 69.45 70.89 50.96 46.22 124.61 128.91 97.88 89.07 907.15 934.34 564.34 511.19 Left BB H 68.57 69.97 50.24 46.71 138.81 139.69 106.74 105.67 1000.8 1115.11 614.11 583.55 III S 71.64 67.65 52.01 43.55 122.97 129.74 95.38 92.52 948.63 930.36 562.36 539.49 Right BB H 68.56 69.63 49.72 44.79 135.94 145.8 107.5 104.28 1000.9 1110.61 612.61 543.35 women with the SEMG device in Karaman State Hospital times with breaks of 90 seconds. In this procedure, ﬁrst the (Figure 7). data given from the hospital system were recorded and then As shown in Table 3, this procedure was carried out the moving RMS was calculated on the analogue and digital through the data of 108 measurements in total, obtained ﬁlter data obtained from the system. through volunteers being unattached, lifting dumbbells of In the system designed as a result of this measurement, 5 kg and the maximum weight they could lift isometrically accuracies of 90.95%± 3.35 for the left BB and 90.75%± 3.75 (1 RM) ﬁrst in the gym, then in the hospital system for three for the right BB were obtained. CH3-left BB as μV with threshold Haar CH3-left BB as μV with Kalman filter 1,000 1,500 2,000 2,500 3,000 3,500 4,000 4,500 5,000 5,500 6,000 6,500 7,000 12 Journal of Healthcare Engineering Table 4: Moving RMS results in gym as training feedback. Muscles and participants Rep1 Rep2 Rep3 Rep4 Rep5 Rep6 Rep7 Rep8 LBB 1 862 798 738 683 782 556 715 741 LBB 2 845 779 852 786 590 812 796 766 LBB 3 757 725 721 560 712 699 645 736 LBB 4 810 841 840 804 828 832 791 830 LBB 5 704 802 651 670 604 354 558 701 LBB 6 387 413 395 354 367 403 381 370 LBB 7 316 328 372 346 377 302 328 319 RBB 1 876 833 811 790 815 846 704 653 RBB 2 823 817 847 834 649 747 621 770 RBB 3 821 793 766 696 566 884 685 785 RBB 4 832 853 856 821 819 808 809 815 RBB 5 815 763 750 753 718 707 725 714 RBB 6 389 422 418 350 371 402 361 378 RBB 7 331 380 365 351 372 348 314 341 1000 1000 900 900 800 800 700 700 600 600 500 500 400 400 300 300 200 200 100 100 0 0 LBB LBB LBB LBB LBB LBB LBB RBB RBB RBB RBB RBB RBB RBB 1 2 3 4 5 6 7 1 2 3 4 5 6 7 Muscles and participants Rep1 Max. rep Min. rep Last rep Figure 8: ADBC results of participants. 3.3. Moving RMS and %MVC Values. During the training, accuracy. As digitally ﬁltered data are compared, it is the volunteers were asked to perform a second training in seen that TCHW method produces better results com- order to obtain the moving RMS values given back to the pared to Kalman ﬁlter. TCHW can soften data as pro- user as feedback. ,e results are presented in Figure 8 and cessable and can also completely ﬁlter out unwanted Table 4 in terms of ease of investigation. signals between muscle contractions. It also eliminates ,us, it can be seen that the system can achieve mini- the distortions in data expressed as artifact. Kalman ﬁlter mum and maximum values of biopotential changes in appears to soften the data but not to be able to completely muscles during training as in [66, 67]. ﬁlter the signal between muscle contractions. Moreover, it is seen that the system can scale the strength obtained Finally, the users were asked to lift 5 kg of dumbbell and the maximum weight they could lift. ,us, the %MVC was as moving RMS during the training on the basis of % MVC with the success rate of 96.87%± 2.74 in terms of measured to be used in performance feedback through the obtained moving RMS values. ,e results obtained are eﬃciency. ,is allows the data obtained to be used in the presented in Table 5. simultaneous performance monitoring and analysis of If Table 5 is analysed, it can be seen that the system can athletes. measure eﬃciency during training with the success rate of 96.87%± 2.74 based on %MVC. 4. Conclusion When data obtained from the designed SEMG system are compared with data obtained from the systems used ,anks to this system, it is thought that athletes will be in the biomedical ﬁeld, it is seen that it has 90.85% able to examine their performances instantly for each M-RMS values (μV) Journal of Healthcare Engineering 13 Table 5: %MVC results in gym. Acknowledgments Muscle name Part. no. kg SMVC (μV) MVC (μV) ,e authors would like to thank Assistant Professor Dr. MVC Yusuf Er (Karamanoglu ˘ Mehmetbey University Physical 5 138.3 850.26 16.26 1 Education and Sports Teaching-Recreation Management) 17.5 845.47 850.26 99.43 for his helpful advice on various technical issues and Atilla 5 138.30 992.6 13.93 Sonmezı ¨ ¸sık (Antalya Sport Center) for his training support. 25 987.15 992.6 99.45 5 152.6 960.13 15,89 20 898.25 960.13 93,55 References 5 155.4 963.30 16.13 LBB 4 20 929.1 963.30 96.44 [1] R. M. Howard, R. Conway, and A. J. Harrison, “Muscle ac- 5 158.9 967.5 16.42 tivity in sprinting: a review,” Sports Biomechanics, vol. 17, 20 956.76 967.5 98.88 no. 1, pp. 1–17, 2018. 5 89.5 615.7 14.53 [2] D. R. Clark, M. I. Lambert, and A. M. Hunter, “Contemporary 15 614.74 615.7 99.84 perspectives of core stability training for dynamic athletic 5 93.5 545.2 17.14 performance: a survey of athletes, coaches, sports science and 12.5 538.9 545.2 98.84 sports medicine practitioners,” Sports Medicine-Open, vol. 4, 5 136.5 875.9 15.58 no. 1, pp. 32–42, 2018. 17.5 854.15 875.9 97.51 [3] J. Hussain, K. Sundaraj, Y. F. Low, L. C. Kiang, S. Sundaraj, 5 140.22 996.4 14.07 and M. A. Ali, “A systematic review on fatigue analysis in 25 985.92 996.4 98.94 triceps brachii using surface electromyography,” Biomedical 5 154.42 972.4 15.88 Signal Processing and Control, vol. 40, pp. 396–414, 2018. 20 893.56 972.4 91.89 [4] R. Lovell, M. Knox, M. Weston, J. C. Siegler, S. Brennan, and 5 156.8 968.7 16.18 P. W. M. Marshall, “Hamstring injury prevention in soccer: RBB 4 20 930.76 968.7 96.08 before or after training?,” Scandinavian Journal of Medicine 5 164.45 972.65 16.90 and Science in Sports, vol. 28, no. 2, pp. 658–666, 2018. 20 910.5 972.65 93.61 [5] S. Skals, J. Vinstrup, E. Sundstrup, M. D. Jakobsen, 5 97.1 614.4 15.80 C. H. Andersen, and L. L. Andersen, “Shoulder and arm 15 609.9 614.4 99.26 muscle activity during elastic band exercises performed in a 5 92.2 515.16 17.89 hospital bed,” De Physician and Sportsmedicine, vol. 46, no. 2, 12.5 506.3 515.16 98.28 pp. 233–241, 2018. [6] M. Cardinale and M. C. Varley, “Wearable training-moni- toring technology: applications, challenges, and opportuni- training and make their training more eﬃcient. It is ties,” International Journal of Sports Physiology and Performance, vol. 12, no. s2, pp. S2–S55, 2017. possible to create intelligent training corners by using the [7] A. D. Vigotsky, I. Halperin, G. J. Lehman, G. S. Trajano, and system in gyms. It is thought that the system can easily be T. M. Vieira, “Interpreting signal amplitudes in surface used by athletes, trainers, kinesiologists, and re- electromyography studies in sport and rehabilitation sci- habilitation experts in bodybuilding trainings and re- ences,” Frontiers in Physiology, vol. 8, pp. 1–15, 2018. habilitation processes. It is possible to improve system [8] E. Zemkova´ and D. Hamar, “Sport-speciﬁc assessment of the features by increasing the number of channels, further eﬀectiveness of neuromuscular training in young athletes,” reducing the PCB size and adding extra sensor. It can be Frontiers in Physiology, vol. 9, pp. 1–27, 2018. possible to follow more complicated movements (deadlift, [9] M. Vromans and P. Faghri, “Electrical stimulation frequency barbell row, etc.) by increasing the number of channels. By and skeletal muscle characteristics: eﬀects on force and fa- making the size of system smaller, it can be possible to tigue,” European Journal of Translational Myology, vol. 27, place it into textile product. In addition, by adding the no. 4, pp. 239–245, 2017. [10] M. Kos and I. Kramberger, “A wearable device and system for pulse oximetry sensor to the system, oxygen consumption movement and biometric data acquisition for sports appli- can be observed during the training. In our future studies, it cations,” IEEE Access, vol. 5, pp. 6411–6420, 2017. is being thought of supporting the system with an image [11] O. Ozkaraca, A. Hakan Isik, and I. Guler, “Biological signals processing system in order to determine movement dis- sensing with wearable technologies and an ECG athlete de- tortions in addition to use it for monitoring training sign,” Recent Patents on Biomedical Engineeringe, vol. 4, no. 2, performance and eﬃciency. pp. 74–83, 2011. [12] R. T. Li, S. R. Kling, M. J. Salata, S. A. Cupp, J. Sheehan, and Data Availability J. E. Voos, “Wearable performance devices in sports medi- cine,” Sports Health: A Multidisciplinary Approach, vol. 8, ,e data used to support the ﬁndings of this study are in- no. 1, pp. 74–78, 2015. cluded within the article. [13] S. Saponara, “Wearable biometric performance measurement system for combat sports,” IEEE Transactions on In- strumentation and Measurement, vol. 66, no. 10, pp. 2545– Conflicts of Interest 2555, 2017. ,e authors declare that there are no conﬂicts of interest [14] I. S. Dhindsa, R. Agarwal, and H. S. Ryait, “Performance regarding the publication of this paper. evaluation of various classiﬁers for predicting knee angle from 14 Journal of Healthcare Engineering electromyography signals,” Expert Systems, vol. 36, no. 3, [29] Delsys, Natick, MA, USA, “Trigno lab,” 2017, http://www. pp. e12381–14, 2019. delsys.com/products/wireless-emg/trigno-lab/. [15] Y. Yin, Y. Zeng, X. Chen, and Y. Fan, “,e internet of things in [30] bitalino, Lisbon, Portugal, “EMG 151015,” 2017, http:// healthcare: an overview,” Journal of Industrial Information bitalino.com/datasheets/EMG_Sensor_Datasheet.pdf/. Integration, vol. 1, pp. 3–13, 2016. [31] Muscle Monitor, Kuopio, FINLAND, “myontec,” 2019, [16] V. Bonaiuto, N. Lanotte, C. Romagnoli, N. Silvaggi, and https://www.myontec.com/musclemonitor/manuals/Old/3_ G. Annino, “Design of a wireless wearable DAQ system for the 1/Muscle%20Monitor%20-%20User%20Guide.pdf12. evaluation of sports performances,” Proceedings, vol. 2, no. 6, [32] Mpower, Vimpeli, Finland, “Mpower,” 2019, http://mpower- pp. 290–295, 2018. bestrong.com/manual/MPOWER_user_manual.pdf. [17] S. Heywood, Y. H. Pua, J. McClelland et al., “Low-cost [33] MyoTrac, Montreal, Quebec, Canada, “MyoTrac3,” 2019, electromyography—validation against a commercial system http://www.thoughttechnology.com/pdf/mar870-00% using both manual and automated activation timing 20myotrac3%20rehab.pdf. thresholds,” Journal of Electromyography and Kinesiology, [34] H. P. Da Silva, “Physiological sensing now open to the world: vol. 42, pp. 74–80, 2018. new resources are allowing us to learn, experiment, and create [18] G. Yang, J. Deng, G. Pang et al., “An IoT-enabled stroke imaginative solutions for biomedical applications,” IEEE rehabilitation system based on smart wearable armband and Pulse, vol. 9, no. 2, pp. 9–11, 2018. machine learning,” IEEE Journal of Translational Engineering [35] A. Burns, B. R. Greene, M. J. McGrath et al., in Health and Medicine, vol. 6, pp. 1–10, 2018. “SHIMMER —A wireless sensor platform for noninvasive [19] A. Ranavolo, G. Chini, A. Silvetti, S. Mari, M. Serrao, and biomedical research,” IEEE Sensors Journal, vol. 10, no. 9, F. Draicchio, “Myoelectric manifestation of muscle fatigue in pp. 1527–1534, 2010. repetitive work detected by means of miniaturized sEMG [36] Viking On Nicolet, Planegg, Germany, “Viking on nicolet,” sensors,” International Journal of Occupational Safety and 2019, https://www.neuroswiss.ch/view/data/5962/Nicolet- Ergonomics, vol. 24, no. 3, pp. 464–474, 2018. Viking-EDX.pdf. [20] A. Cisnal, R. Alonso, J. P. Turiel, J. C. Fraile, V. Lobo, and [37] K. Biron, K. Englehart, and P. Parker, “A study of kalman V. Moreno, “EMG based bio-cooperative direct force control of ﬁltering applied to myoelectric signal state tracking,” CMBES, an exoskeleton for hand rehabilitation: a preliminary study,” in vol. 35, pp. 1–4, 2012. Converging Clinical and Engineering Research on Neuro- [38] T. Boaventura, L. Hammer, and J. Buchli, “Interaction force rehabilitation III. ICNR 2018. Biosystems & Biorobotics, estimation for transparency control on wearable robots using L. Masia, S. Micera, M. Akay, and J. Pons, Eds., vol. 21, Springer, a kalman ﬁlter,” in Converging Clinical and Engineering Re- Cham, Switzerland, 2019. search on Neurorehabilitation II. Biosystems & Biorobotics, [21] T. D. Williams, D. V. Tolusso, M. V. Fedewa, and M. R. Esco, J. Ibañez, ´ J. Gonzalez-Vargas, ´ J. Azor´ın, M. Akay, and J. Pons, “Comparison of periodized and non-periodized resistance Eds., vol. 15, Springer, Cham, Switzerland, 2017. training on maximal strength: a meta-analysis,” Sports [39] B. J. Schoenfeld, “,e mechanisms of muscle hypertrophy and Medicine, vol. 47, no. 10, pp. 2083–2100, 2017. their application to resistance training,” Journal of Strength [22] R. A. Panariello, T. J. Stump, and F. A. Cordasco, “,e lower and Conditioning Research, vol. 24, no. 10, pp. 2857–2872, extremity athlete: postrehabilitation performance and injury prevention training,” Operative Techniques in Sports Medi- [40] B. J. Wallace, H. C. Bergstrom, and T. A. Butterﬁeld, cine, vol. 25, no. 3, pp. 231–240, 2017. “Muscular bases and mechanisms of variable resistance [23] J. Gondin, P. J. Cozzone, and D. Bendahan, “Is high-frequency training eﬃcacy,” International Journal of Sports Science and neuromuscular electrical stimulation a suitable tool for Coaching, vol. 13, no. 6, pp. 1177–1188, 2018. muscle performance improvement in both healthy humans [41] E. Barreiro, “Models of disuse muscle atrophy: therapeutic and athletes?,” European Journal of Applied Physiology, implications in critically ill patients,” Annals of Translational vol. 111, no. 10, pp. 2473–2487, 2011. Medicine, vol. 6, no. 2, p. 29, 2018. [24] S. Grosprˆetre, P. Gimenez, L. Mourot, and G. Coratella, [42] J. Kilby, K. Prasad, and G. Mawston, “Multi-channel surface “Elastic band exercise induces greater neuromuscular fatigue electromyography electrodes: a review,” IEEE Sensors Journal, than phasic isometric contractions,” Journal of Electromyog- vol. 16, no. 14, pp. 5510–5519, 2016. raphy and Kinesiology, vol. 47, pp. 113–120, 2019. [43] D. Devaprakash, G. J. Weir, J. J. Dunne, J. A. Alderson, and [25] S. C. F. A. Von Werder, T. Kleiber, and C. Disselhorst-Klug, C. J. Donnelly, “,e inﬂuence of digital ﬁlter type, amplitude “A method for a categorized and probabilistic analysis of the normalisation method, and co-contraction algorithm on surface electromyogram in dynamic contractions,” Frontiers clinically relevant surface electromyography data during in Physiology, vol. 6, no. 30, pp. 1–8, 2015. clinical movement assessments,” Journal of Electromyography [26] U. Imtiaz, L. Bartolomeo, Z. Lin et al., “Design of a wireless and Kinesiology, vol. 31, pp. 126–135, 2016. miniature low cost EMG sensor using gold plated dry elec- [44] Texas instrument, May 2018, http://www.ti.com/lit/ds/symlink/ trodes for biomechanics research,” in Proceedings of the 2013 ina128.pdf. IEEE International Conference on Mechatronics and Auto- [45] Texas instrument, June 2018, http://www.ti.com/lit/ds/symlink/ mation, pp. 957–962, Takamatsu, Japan, August 2013. tl072.pdf. [27] G. R. Jones, N. A. Neubauer, B. O’Connor, and J. M. Jakobi, ´ ´ ´ ´ [46] N. Miljkovic, N. Popovic, O. Djordjevic, L. Konstantinovic, “EMG Functional tasks recordings determines frailty phe- and T. B. Sekara, “ECG artifact cancellation in surface EMG notypes in males and females,” Experimental Gerontology, signals by fractional order calculus application,” Computer vol. 77, pp. 12–18, 2016. [28] S. Sadeghi Esfahlani, B. Muresan, A. Sanaei, and G. Wilson, Methods and Programs in Biomedicine, vol. 140, pp. 259–264, “Validity of the Kinect and Myo armband in a serious game 2017. for assessing upper limb movement,” Entertainment Com- [47] Analog devices, September 2018, https://www.analog.com/en/ puting, vol. 27, pp. 150–156, 2018. analog-dialogue/articles/practical-ﬁlter-design-precision-adcs.html. Journal of Healthcare Engineering 15 [48] A. Widmann, E. Schroger, ¨ and B. Maess, “Digital ﬁlter design voluntary contractions of human adductor pollicis,” Journal of Biomechanics, vol. 40, no. 1, pp. 1–8, 2007. for electrophysiological data - a practical approach,” Journal of Neuroscience Methods, vol. 250, pp. 34–46, 2015. [64] D. Roman-Liu and P. Bartuzi, “Inﬂuence of type of MVC test on electromyography measures of biceps brachii and triceps [49] A. F. T. Ibrahim, V. R. Gannapathy, L. W. Chong, and I. S. M. Isa, “Analysis of electromyography (EMG) signal for brachii,” International Journal of Occupational Safety and Ergonomics, vol. 24, no. 2, pp. 200–206, 2018. human arm muscle: a review,” in Lecture Notes in Electrical [65] J. J. D. Van der Gon, B. M. ter Haar Romeny, and Engineering, pp. 567–575, 2016. E. J. van Zuylen, “Behaviour of motor units of human arm [50] M. B. I. Raez, M. S. Hussain, and F. Mohd-Yasin, “Techniques muscles: diﬀerences between slow isometric contraction and of EMG signal analysis: detection, processing, classiﬁcation relaxation,” De Journal of Physiology, vol. 359, no. 1, and applications,” Biological Procedures Online, vol. 8, p. 11, pp. 107–118, 1985. [66] R. U. Newton, W. J. Kraemer, K. Hakkinen, ¨ B. J. Humphries, [51] SENIAM, March 2017, http://www.seniam.org/. and A. J. Murphy, “Kinematics, kinetics, and muscle acti- [52] F. Negro, K. Keenan, and D. Farina, “Power spectrum of the vation during explosive upper body movements,” Journal of rectiﬁed EMG: when and why is rectiﬁcation beneﬁcial for Applied Biomechanics, vol. 12, no. 1, pp. 31–43, 1996. identifying neural connectivity?,” Journal of Neural Engi- [67] S. Maeo, Y. Yoshitake, Y. Takai, T. Fukunaga, and neering, vol. 12, no. 3, article 036008, 2015. H. Kanehisa, “Neuromuscular adaptations following 12-week [53] K. Praveen, A. D. Arun, and M. Sivaramakrishna, “EMI/EMC maximal voluntary co-contraction training,” European qualiﬁcation of diﬀerential pressure monitoring system for Journal of Applied Physiology, vol. 114, no. 4, pp. 663–673, PFBR—A case study,” in Proceedings of the 2016 International Conference on ElectroMagnetic Interference & Compatibility (INCEMIC), pp. 1–4, Bangalore, India, December 2016. [54] J. Zhang, L. Wang, L. Yu, Y. Yang, Y. Zhang, and B. Li, “A low- oﬀset analogue front-end IC for multi-channel physiological signal acquisition,” in Proceedings of the 2009 Annual In- ternational Conference of the IEEE Engineering in Medicine and Biology Society, pp. 4473–4476, Minneapolis, MN, USA, September 2009. [55] C. K. Chui and G. Chen, “Kalman ﬁlter: an elementary ap- proach,” in Kalman Filtering, Springer, Cham, Switzerland, [56] A. Toprak and I. Guler, ¨ “Angiograph image restoration with the use of rule base fuzzy 2D Kalman ﬁlter,” Expert Systems with Applications, vol. 35, no. 4, pp. 1752–1761, 2008. [57] M. Schimmack, S. Nguyen, and P. Mercorelli, “Anatomy of haar wavelet ﬁlter and its implementation for signal pro- cessing,” IFAC-PapersOnLine, vol. 49, no. 6, pp. 99–104, 2016. [58] S. Roldan-Vasco, E. Perez-Giraldo, and A. Orozco-Duque, “Continuous wavelet transform for muscle activity detection in surface EMG signals during swallowing,” in Applied Computer Sciences in Engineering. WEA 2018. Communica- tions in Computer and Information Science, J. Figueroa- Garc´ıa, J. Villegas, J. Orozco-Arroyave, and P. Maya Duque, Eds., vol. 916, Springer, Cham, Switzerland, 2018. [59] S. Ranaldi, C. De Marchis, and S. Conforto, “An automatic, adaptive, information-based algorithm for the extraction of the sEMG envelope,” Journal of Electromyography and Ki- nesiology, vol. 42, pp. 1–9, 2018. [60] S. Mazumdar, A. Saikia, N. Sahai, S. Paul, and D. Bhatia, “Determination of signiﬁcant muscle in movement of upper limb using maximum voluntary contraction of EMG signal,” in Proceedings of the 2017 4th International Conference on Signal Processing and Integrated Networks (SPIN), pp. 96–99, Noida, India, February 2017. [61] B. N. Cahyadi, I. Zunaidi, S. A. Bakar et al., “Upper limb muscle strength analysis for movement sequence based on maximum voluntary contraction using EMG signal,” in Proceedings of the 2018 International Conference on Com- putational Approach in Smart Systems Design and Applica- tions (ICASSDA), pp. 1–5, Kuching, Malaysia, August 2018. [62] J. Paquin and G. A. Power, “History dependence of the EMG- torque relationship,” Journal of Electromyography and Kine- siology, vol. 41, pp. 109–115, 2018. [63] E. N. Rousanoglou, A. E. Oskouei, and W. Herzog, “Force depression following muscle shortening in sub-maximal International Journal of Advances in Rotating Machinery Multimedia Journal of The Scientific Journal of Engineering World Journal Sensors Hindawi Hindawi Publishing Corporation Hindawi Hindawi Hindawi Hindawi www.hindawi.com Volume 2018 http://www www.hindawi.com .hindawi.com V Volume 2018 olume 2013 www.hindawi.com Volume 2018 www.hindawi.com Volume 2018 www.hindawi.com Volume 2018 Journal of Control Science and Engineering Advances in Civil Engineering Hindawi Hindawi www.hindawi.com Volume 2018 www.hindawi.com Volume 2018 Submit your manuscripts at www.hindawi.com Journal of Journal of Electrical and Computer Robotics Engineering Hindawi Hindawi www.hindawi.com Volume 2018 www.hindawi.com Volume 2018 VLSI Design Advances in OptoElectronics International Journal of Modelling & Aerospace International Journal of Simulation Navigation and in Engineering Engineering Observation Hindawi Hindawi Hindawi Hindawi Volume 2018 Volume 2018 Hindawi www.hindawi.com Volume 2018 www.hindawi.com Volume 2018 www.hindawi.com www.hindawi.com www.hindawi.com Volume 2018 International Journal of Active and Passive International Journal of Antennas and Advances in Chemical Engineering Propagation Electronic Components Shock and Vibration Acoustics and Vibration Hindawi Hindawi Hindawi Hindawi Hindawi www.hindawi.com Volume 2018 www.hindawi.com Volume 2018 www.hindawi.com Volume 2018 www.hindawi.com Volume 2018 www.hindawi.com Volume 2018

Journal

Journal of Healthcare Engineering – Hindawi Publishing Corporation

Published: Sep 9, 2019

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

Learn More →

Design and Validation of Multichannel Wireless Wearable SEMG System for Real-Time Training Performance Monitoring

Design and Validation of Multichannel Wireless Wearable SEMG System for Real-Time Training Performance Monitoring

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

Learn More →

Design and Validation of Multichannel Wireless Wearable SEMG System for Real-Time Training Performance Monitoring

Design and Validation of Multichannel Wireless Wearable SEMG System for Real-Time Training Performance Monitoring

References (71)

Abstract

Journal

Recommended Articles

There are no references for this article.

Our policy towards the use of cookies