Access the full text.
Sign up today, get DeepDyve free for 14 days.
Loading next page...
/lp/multidisciplinary-digital-publishing-institute/risk-assessment-for-uas-logistic-delivery-under-uas-traffic-management-WHBdoPT36I
References
References for this paper are not available at this time. We will be adding them shortly, thank you for your patience.
- Publisher
- Multidisciplinary Digital Publishing Institute
- Copyright
- © 1996-2020 MDPI (Basel, Switzerland) unless otherwise stated Disclaimer The statements, opinions and data contained in the journals are solely those of the individual authors and contributors and not of the publisher and the editor(s). Terms and Conditions Privacy Policy
- ISSN
- 2226-4310
- DOI
- 10.3390/aerospace7100140
- Publisher site
- See Article on Publisher Site
Abstract
aerospace Article Risk Assessment for UAS Logistic Delivery under UAS Trac Management Environment Pei-Chi Shao Department of Aviation & Maritime Transportation Management, Chang Jung Christian University, Tainan 711301, Taiwan; pcshao@mail.cjcu.edu.tw Received: 25 August 2020; Accepted: 19 September 2020; Published: 25 September 2020 Abstract: Resulting from a mature accomplishment of the unmanned aircraft system (UAS), it is feasible to be adopted into logistic delivery services. the supporting technologies should be identified and examined, accompanying with the risk assessment. This paper surveys the risk assessment studies for UAVs. the expected level of safety (ELS) analysis is a key factor to safety concerns. By introducing the UTM infrastructure, the UAS implementation can be monitored. From the NASA technical capability level (TCL), UAV in beyond visual line of sight (BVLOS) flights would need certain verifications. Two UAS logistic delivery case studies are tested to assert the UAS services. To examine the ELS to ground risk and air risk, the case studies result in acceptable data to support the UAS logistic delivery with adequate path planning in the remote and suburban areas in Taiwan. Keywords: UAS safety; risk assessment; expected level of safety (ELS); UAV logistic delivery 1. Introduction The advances of micro electro-mechanical systems (MEMS) have led the unmanned aircraft system (UAS) into rapid growth since 2012. the vertical take-o and landing (VTOL) multi-rotor unmanned aerial vehicle (UAV) has made very successful progress in characteristics of simple structure, easy operation, and good performance to suit for wide applications. Thus, UAVs can easily be used to work with dirty, dangerous, and dull (3D) jobs with higher operational eciency and personnel safety. VTOL UAVs have also successfully grasped consumer markets, and have replaced many conventional systems into new visions. The International Civil Aviation Organization (ICAO) document 328 for unmanned aircraft systems (UASs) gives the definition of UAS as “An aircraft and its associated elements which are operated with no pilot on-board.” VTOL UAVs use brushless (BL) motors, electronic speed controllers (ESC), Li polymer (LiPo) battery, GPS navigation, and inertia navigation microcontroller to build on fiber materials with a required payload through the radio link to a ground controller to fly. This creates a friendly UAS operation environment [1,2]. The UAV flight operation and management in Taiwan is legislated by the Ministry of Transportation and Communications (MOTC), to approve the legal use of UAVs in non-integrated airspace (NIA) by the “Civil Aviation Act”. This was legislated on 3 April 2018 [3] and is e ective on 31 March 2020. This UAV law regulates UAS operations and is enforced into surveillance and management to assure the UAS flight safety. Research and development on the UAS trac management (UTM) with appropriate UAS business models and applications are booming. Several UAS services and business models are established for journalist/news broadcasting air photography, trac surveillance, agriculture insecticide spray, mountainside slide inspection, bridge inspection, and logistic delivery, etc. The Federal Aviation Administration (FAA) advisory circular (AC) no. 107-2 states that the operational limitations for small UAV (sUAV) are to fly less than a ground speed of 160 km/h and lower than 400 feet above ground level (AGL) [4]. AC 107-2 by FAA allows all the applications of sUAV Aerospace 2020, 7, 140; doi:10.3390/aerospace7100140 www.mdpi.com/journal/aerospace Aerospace 2020, 7, x FOR PEER REVIEW 2 of 19 The Federal Aviation Administration (FAA) advisory circular (AC) no. 107-2 states that the Aerospace 2020, 7, 140 2 of 20 operational limitations for small UAV (sUAV) are to fly less than a ground speed of 160 km/h and lower than 400 feet above ground level (AGL) [4]. AC 107-2 by FAA allows all the applications of to be legitimated for delivery of goods, surveillance, and search and rescue by Kopardekar et al. [5]. sUAV to be legitimated for delivery of goods, surveillance, and search and rescue by Kopardekar et Figure 1 shows a typical commercial legal application involving Amazon’s logistics trials using sUAVs al. [5]. Figure 1 shows a typical commercial legal application involving Amazon’s logistics trials using in NIA for air trac control (ATC) under 400 feet AGL [6,7]. the European Union Aviation Safety sUAVs in NIA for air traffic control (ATC) under 400 feet AGL [6,7]. The European Union Aviation Agency (EASA) and Single European Sky ATM Research (SESAR) also define the U-space concept for Safety Agency (EASA) and Single European Sky ATM Research (SESAR) also define the U-space sUAVs NIA below 700 feet [8,9]. Di erent regulations from FAA and EASA can be adopted to suit for concept for sUAVs NIA below 700 feet [8,9]. Different regulations from FAA and EASA can be UAV operations to di erent countries and territories. adopted to suit for UAV operations to different countries and territories. Figure 1. Amazon’s non-integrated airspace for unmanned aerial vehicles (UAVs) [6]. Figure 1. Amazon’s non-integrated airspace for unmanned aerial vehicles (UAVs) [6]. In In T Tai aiwan, wan, the the NIA NIA below below 400 400 feet feet is is authorized authorized and and commissioned commissioned to to local local governments governments for for management; while higher altitude flights are controlled by the CAA authority and air trac control management; while higher altitude flights are controlled by the CAA authority and air traffic control (A (ATC) TC) [[3 3] ].. Under Under such such cir circ cumstance, umstance, a a hierar hierarchical chical UTM UTM is is designed designed and and constr construct ucted ed to to include include all all UAVs into regional UTM (RUTM) or national UTM (NUTM) [10,11]. the hierarchical UTM proposes UAVs into regional UTM (RUTM) or national UTM (NUTM) [10,11]. The hierarchical UTM proposes an an ADS-B ADS-B like like infrastr infrastruct uctur ure e with with an an on-boar on-board d unit unit (OBU) (OBU) to to br bro oadcast adcast flight flight data data down down to to gr ground ound transceiver stations (GTS) to UTM cloud for surveillance. the surveillance data include position track transceiver stations (GTS) to UTM cloud for surveillance. The surveillance data include position track and and flight flight data data with with six-degr six-degr ees-of-fr ees-of eedom, -freedom to transpar , to transp ently ar watch ently all watch UAVs all flying UAVs in the flying responsible in the airspace. the preliminary tests verify the importance of introducing UTM for UAV surveillance [10]. responsible airspace. The preliminary tests verify the importance of introducing UTM for UAV surve Due illance to the [10]. lack of statistical data for the past UAS flight experience, it is dicult to collect real flight data for risk analysis. From the literature survey, many studies adopted general aviation (GA) data Due to the lack of statistical data for the past UAS flight experience, it is difficult to collect real to flight simulate data fUA or risk a V risk na assessment lysis. From [th 12e ].liHowever terature s ,urvey UAVs, ar man e indeed y studie incr s ad easingly opted ge used nerafor l aviation commer (GA cial ) or private activities, further studies in deep involvement become urgent and critical. the developing data to simulate UAV risk assessment [12]. However, UAVs are indeed increasingly used for UTM commerc system ial otries r priv to ate collect activities, UAVfbig urther data studies from UA in d Vee surveillance p involvemwith ent beco flight me data urgent for and flight croperation itical. The quality developassurance ing UTM s (FOQA) ystem tranalysis ies to col [10 lect ]. UAV UAS FOQA big datcan a from support UAV UA sur V ve risk illance analysis with with flightr eal data data. for the flight safe op operation eration quality of sUAas Vs surance becomes (FO a gr QA) eat an concern alysis to [10]. theUA public S FO and QA turns can sinto uppo art significant UAV risk challenge analysis to with aviation real dat safety a. Th .e Studies safe operation includeof risk sUAVs assessment becomes concerning a great con airspace, cern to th UA e p V ubli MRO c and (maintenance, turns into a r si epair gnifica , and nt over challenge haul) in to manipulation aviation safet[y. 1,13 Studies ]. To investigate include risk theas UAS sessrisk menand t con safety cerning , a logistic airspace delivery , UAV integrated MRO (mainte pilot nance, program repa(IPP) ir, and is demonstrated overhaul) in man for ipu technical lation verification [1,13]. To investig in this at paper e the . UAS risk and safety, a On l the ogist UA ic Vdel risk ivery assessment, integrated it p isiltermed ot progr into am (I air PP) risk is demonst and ground rated for risk classifications technical verifica [14,15 tion ]. In in the thisUA paper V air . risk, collision avoidance resulting from geofensing, detect and avoid (DAA) mechanism, and fr On ontth detect e UAV is ris focused k assessme on. While nt, it in is term the UA ed V into ground air ririsk, sk an crash d ground mode ris [11 k ], classifi ground catimpact, ions [14,1 kinetic 5]. In ener the UAV gy, and air debris risk, col casualty lision avoid [16,17 ance ] due resto ulting failur from es wer geofens e analyzed ing, dein tec pr t and ofundity avoid and (DAhave A) mec been hanism paid, highly and front attentions detect is in foc the used UAS onoperation. . While in the UAV ground risk, crash mode [11], ground impact, kinetic energy A typical , and debris risk assessment casualty [16, study 17] focusing due to failures on sUA we Vsre was anundertaken alyzed in pro by fu the ndity FAA and [1,12 have ] to determine been paid the high risk ly atten leveltions in of flying thsUA e UA Vs S op over eration. di er ent types of areas. In this report, only small UAVs of less than 250 gA and typ flying ical risk at less asses than sment 25 m study /s were focu taken sing into on account. sUAVs was However unde , rtaken in reality by , ath velocity e FAA of [1,12 25 m ] to /s can determ only ine be th achieved e risk level in of fixed flyin wings g sUA[Vs 1]. ov Frer om different real experience, types of a multi reas. In rotor this UA repo Vs rt, ar onl e flying y sma4~8 ll UAVs m/s with of less their than maximum 250 g and take-o flying at weight less than (MT 25 OW) m/s were around tak10 en kg into that acco may untbe . Howev a typical er, in carrier reality, for a delivery velocity . In of 25 addition, m/s can wind only e be ect achie may ved interr in fixed upt wings sUAS [flight 1]. Fro operations m real experi due ence, to the mul small ti roto flight r UAVs momentum. are flying the 4~8 m flight /s wi conditions th their m need aximverifications um take-off we by ig real ht flights. (MTOW Logistic ) around delivery 10 kg th flight at may be cases will a typ be ical discussed carrier for in this paper. Aerospace 2020, 7, 140 3 of 20 This paper accounts an UAS risk assessment study using flight scenarios that involves the FAA formulation for commercial UAVs of 12 kg MTOW flying at 5–8 m/s [11]. the logistic delivery case study is used to support the analysis of risk assessment and the expected level of safety (ELS) and for further estimation on risk level to population density in remote and suburban areas. UAV accidents are monitored by the aviation authority and police administrations. the Civil Aviation Act legislates regulations for UAV flights. UAVs can plan routine flights in legal airspaces, which are colored as yellow and green areas, except the restricted red areas in Taiwan [3]. the operation of UAVs brings a significant facilitation of convenience but impacts public safety. Before 2004, without adequate regulations, UAVs were governed by the following guidelines [13]: UAV Operations Shall not Increase the Risk to Other Airspace Users or Third Parties The Joint Aviation Authorities (JAA) used an “Equivalent Risk” for UAV operations [13] to monitor the safety of UAVs and UAV accidents for the past 15 years. Due to the lack of UAV event/accident data, manned aircraft event/accident data have been referred to determine the risk level [13,18] for UAVs. In order to determine the risk of UAV operations, some safety factors are proposed using the linear model, such as the physical factors of weight, velocity, kinetic energy (KE) and frontal impact area [11], the ground population, and the e ect of shelter and the number of casualties [18,19]. Frequent UAV activities in the near future imply that air trac monitor and control in the low-altitude airspace is required with an e ective methodology in surveillance [5,10]. “Specific Operations Risk Assessment” (SORA) by the Joint Authorities for rulemaking of unmanned system (JARUS) [15] claims that the UAS risk is a combination of probability of any associated levels of severity in occurrence. the safety level defined by probability fatalities is classified on the ground or in the air. the SORA provides a systematic methodology to identify risks associated with an UAS operation in a holistic way. the SORA process is a valuable tool to standardize the risk framework in UAS operations. For example, in order to reduce the air risk for mid-air collision, the SORA takes the tactical mitigations by the DAA mechanism or alternate means of services of operational procedures. Since DAA is one of the ways to mitigate the air risk in mid-air collision, the hierarchical UTM system can detect UAVs in the approach and command and avoidance by software manipulation. With real data performance through real time flight surveillance, software DAA is achieved [10]. a principal task of UTM is applied to accurately determine the safety level of UAVs. An UAV delivery is demonstrated as an integrated pilot program (IPP) for UAS under UTM surveillance. It is used to verify the feasible use of UAS in logistic delivery. From the results, it does strongly enable the risk level assessment from such UAS applications. A hexa-rotor VTOL UAV is tested in a remote area and suburban area for risk assessment under UTM surveillance. the purpose of the tests is to verify the technical feasibility and capability. the quantitative analysis is not focused on in this phase of experiments. the flights carried 2.4 kg flying at 35 m above ground level (AGL) for delivery over 5 km away. the demonstrations were successful in terms of autopilot performance flying beyond the visual line of sight (BVLOS). To determine the safety use of UAV in logistic delivery, the risk analysis for this case study has arisen to take flight planning into account [11]. Although the mean time between failures (MTBF) for VTOL UAVs is still as low as 100 h, the risk assessment for this study results in a safety measure to banish society concerns to UAS operations. 2. Risk Orientation The UAV risk assessment in terms of safety management uses the pilot experiences and safety records for manned aircraft [12]. However, UAVs carry no passengers, so the safety standard for UAV can be focused on the protection of third parties and property. To increase public safety, an equivalent level of safety (ELOS) for manned aircraft was used to certify UASs by the Joint Aviation Authority (JAA) in 2004 [12,20]. Weibel and Hansman [17] demonstrated the use of ELOS to determine the operational Aerospace 2020, 7, 140 4 of 20 requirements for di erent classes. However, it is dicult to quantify an ELOS for UAV applications due to the lack of real data. Dalamagkidis et al. used the standard components of the system to establish a target level of safety (TLS) [21]. Neither ELOS nor TLS can be used for a real UAV safe assessment using the simulated data. an equivalent analysis concept is used to determine the similarities in risk development for UAVs and manned aircraft. UAV technologies are elevating into maturity rapidly. Runaway pilots operating incontrollable UAVs have threatened aviation safety near airports in Taiwan since 2015 [3]. an orderly regulated UAV surveillance under UTM is expected. Recent studies of UAV risk relate to the sUAV operational safety in low-altitude airspace focusing on UTM [22], ground impact hazard [17], the third party casualty risk [1,13], and the target level of safety (TLS) [18,21]. Under scheduled IPP flight tests, the UAS operation safety or risk can be analyzed and estimated from the observable and controllable data. 2.1. Ground Risk Assessment The SORA [16] describes the guidelines to approach safety created by the UAS operations for specific assurance and integrity levels (SAIL) into either ground risk and air risk. Both risk and SAIL can be reduced by the e ective methodology. It can be accomplished through the UAS operators by utilizing certain threat barriers and mitigating measures. From which, UTM is one of legal, feasible, and e ective solutions. Under such understanding, risk assessment and prevention from ground and air shall be taken into account seriously. Most studies refer to the general aviation or air transport aircraft. However, this is indeed impractical. In this study, the risk assessment study considers some preliminary tests with certain scenarios to assert a feasible evaluation. The event tree analysis was used to analyze four scenarios for harm to the public on the ground from the e ect of UAV operations to impact public safety. the events include: (1) Failure of UAS, (2) impact in the populated area, (3) debris penetration to sheltering, and (4) resulting fatal penetration [13,16,17]. Lin and Shao [10] analyzed the UAV air crash behavior to explore the severity of ground impact resulting from the experiments. the UAV ground impact event tree shows the risk factors for UAV failure. the “Ground Impact Hazard” is a model to determine the e ect of di erent factors on the expected level of safety (ELS) [17,23,24]. It is expressed in terms of ground fatality events per hour of flight (E/h). the factors for ground impact include the total system reliability, UAV size, UAV kinetic energy (KE) at the moment of power loss, and population density near UAV flight operations [17]. Since there is very little data about UAV fatalities [15], the TLS for the ground impact model uses a value of 1 10 E/h, which is recommended by the FAA for manned aircraft operations. In the reference, the TLS for air transportation ground fatalities is 2 10 E/h, which is recommended by the National Transportation Safety Board (NTSB) database. These data can fit either manned aircraft or unmanned vehicles. According to these data, the ELS for the ground impact model for an UAV is 1 10 E/h fatalities [17]. Melnyk et al. [13] used a linear model to estimate casualties and determine the UAS risk. the model parameters include area population, shelter e ect, and frontal impact area casualties. In terms of the linear model, the parameters for UAVs are frontal impact area [10], kinetic energy, and the e ect of shelter [1,20]. There is a wide range of UAS sizes and characteristics, such as the on-board flight control system and/or the presence of a communication link. Therefore, the model for manned aircraft ELOS cannot directly apply to UAVs because an UAV does not carry passengers and crew. the probability of injuries and fatalities for UAVs are lower than that of any kind of manned aircraft [21]. Since the various accident types have resulted in di erent e ects on the safety level, the ELOS, TLS, and ELS must be categorized for UAV operations of di erent types, power, weight, and altitudes using similar accident scenarios for risk analysis. This is an important key to UAS. an UAS database must be established to assert risk assessment from the causes of failure to reduce the estimation error. a preliminary study on the construction of an UTM by Lin et al. [10] acquired flight control data on the UAV performance for flight operation quality assurance (FOQA). Aerospace 2020, 7, 140 5 of 20 The SORA methodology by JARUS [16] provides the ground risk class (GRC) determination to identify the ground risk level using the risk scores of UAS categorizations of operation. the ground risk level determined by the (1) maximum UAS characteristics dimension from 1, 3, 8 m or larger, (2) their corresponding typical kinetic energy from 700 J, 34 kJ, 1084 kJ or larger, and (3) operational scenarios, such as VLOS, BVLOS, and population area, their intrinsic UAS ground risk classes vary from low to high. Each di erent condition of class impacts the severity of ground risk. It is obvious that the UAS kinetic energy causes a dominant casualty to the ground, either to human life or property. Lin and Shao [11] used the Weibel’s ground impact model [17] to define possible injuries and fatalities, and ELS is the risk assessment tool to identify the risk level. the simulation resulted from an e ective path planning for least crash probability density (CPD) [11]. To determine the damage due to the UAV failure, the physical characteristics of the UAV are used from the FAA in experiments. the most significant factors involve UAV weight and kinetic energy [1,10]. the FAA task force used information from a study by the United Kingdom Ministry of Defense in 2010, which states that an object with a kinetic energy level of 80 J has a 30% probability of striking the head of a person [1,5]. In terms of the number of casualties for a threshold value for mass and velocity, this equates to an object weighing 250 g traveling at a terminal velocity of 25 m/s (approximately 57 m/h). Although the terminal velocity of 25 m/s used in this report is not realistic in experience, the impact harm and injury to people on the ground resulted from failures. In order to determine the probability of a catastrophic event involving an UAV, the FAA task force uses the mean time between failures (MTBF), the population density and the exposed fraction and probability lethality to calculate the probability of an UAV event that results in casualties. the results of experiments by the task force mean that FAA does not require the registration of any UAV with a MTOW < 250 g (0.55 pounds) [4]. Based on the UAV flight path risk, Lin and Shao [11] developed the crash probability density (CPD) radius of UAV path planning. the ELS can provide risk level information to avoid the high population along the path planning in the pre-flight process. In medium- and high-risk environments, UAS applications for logistic delivery over populated areas are allowed in the near future. the European Union Aviation Safety Agency (EASA) releases a special condition to certify UAVs for logistic services. It is open to the public for comments before 30 September 2020 [25]. the proposed certification approach, SC-light UAS, will apply to all UAVs with a MTOW < 600 k (1322 pounds). However, this does not allow transporting passengers in any way, which is operated without a remote pilot being able to intervene. the document also applies to UAV operations in terms of “specific”. the EUs risk-based framework is defined as open, specific, and certified. Rulemaking on the certified category is on-going [25]. In the near future, the urban air mobility (UAM) vehicles may become a di erent story to discuss. 2.2. Air Risk Assessment The NASA UTM technical capability level (TCL) is surveyed based on the four metrics about ground and air risks with UAS operations: (1) Population density, (2) the amount of people and property on the ground, (3) the number of manned aircraft in close proximity to the sUAS operations, and (4) the density of the UAS operations [26]. the TCL verifies and confirms the maturity of UAVs in urban and suburban services, especially logistic delivery. For the air risk management in NAS airspace, Melnyk et al. [18] proposed a framework to develop an e ectiveness standard of sense and avoid (SAA) for UAS. “E ectiveness” is defined as the combination of reliability and ecacy, which also indicates UAS failures or insucient performance standards. In order to develop the minimum e ectiveness standard for SAA, a framework is utilized to include a target level of safety (TLS) approach to the problem and an event tree format risk model to predict mid-air collision (MAC) fatality rates resulting from UAS operations. the event tree model is a risk mitigation by proper separation or collision avoidance. the event tree is a series of branches with Aerospace 2020, 7, 140 6 of 20 options for the air environment, mitigation, and event outcomes based on the probability of each of the branches in occurring and the e ects of progress along a branch [18]. Since UAS needs a vehicle function certification to ensure or minimize with the least in-flight collision and ground impact fatality, this is defined as target levels of safety (TLS). Therefore, Schrage developed the sUAS operators that need a functional safety management (FSM) approach that is a ordable to ensure safety for their limited operations. the air risks are including the functions or subsystems of UAS failure. the risk assessment tools for functional hazard assessment (FHA) and operational risk assessment (ORA) are utilized to complete the UAS safety assessment process in the UAS logistic delivery experiment [27]. FHA identifies and evaluates the hazards associated with functions of the operation system, while ORA evaluates the overall risks associated with each hazard of function. TLS is used as a constraint for sUAS functional safety management (FSM). the key purpose of FHA identifies and evaluates the hazards associated with aircraft-level functions. Based on the research results, the functions associated with failures will have the highest risk dealing with flight control and SAA capabilities. the functional decompositions of sUAS for logistic delivery referring to di erent risk levels are constructed in block diagrams. From this block diagram, SAA is the most significant area for risk mitigation as important as guidance, navigation, and control (GNC) in the flight control [27]. Martin et al. addressed [28] that SORA adopts a holistic view for managing air risk, incorporating greater flexibility. This indicates how mitigation can be combined with a strategic or tactical way. the flexibility deals with a qualitative set of rules on trac density with a continuous performance function of detect, decide, and avoid. the SORA employs three air risk classifications (ARC), where ARC-b, c, d are necessary equipment integrity and assurance requirements. ARC is a qualitative classification for the rate at which an UAS would encounter manned aircraft in NAS. It is an initial hypothesis for aggregated collision risk in the airspace, before any mitigations may be applied. Allouch et al. [14] followed the ISO 12100 to approach risk assessment and risk mitigation by three step risk analyses: (1) First, to start with system limits specification in five categories on physical, temporal, environmental, behavioral limits, and networking limits. (2) Second, performing hazard identification to provide a list of potential drone hazards according to their external and internal sources. (3) Third, estimating UAV risk measures of probabilities and severity levels of the consequences of the identified UAS operational hazards. According to the ISO 12100 standard, the risk estimation consists of determinations on the risk severity and probability. the risk severity is estimated based on the injury level or the harmful impact on people, environment, and UAV itself. the risk severity of the hazard is usually a ected by the degrees of consequence as catastrophic, critical, marginal, and negligible; while the risk probability is recognized by frequent, probable, occasional, remote, and improbable occurrence. the hazard sources from internal and external a ecting factors are important to UAV pilots to prevent from malfunction and failure. The 4G/LTE communication is selected as one possible implementation of information exchange by Allouch et al. [14]. the procedures are specified into pre-flight, in-flight, and post-flight. Each phase needs to establish a standard operation procedure (SOP) to assure that the pilot, vehicle, and environment are being ready and suitable to perform an UAS mission flight. In recent research, Lin and Shao [10] demonstrated the ADS-B like infrastructure to establish a surveillance down link into the UTM. From the literature reviews, UAV operations have di erent levels of risk including flight procedure, system infrastructure, CNS, meteorological and environmental factor, and human factor. Di erent types of UAVs with various specifications may result in di erent levels of risk assessment referring to methodologies and suited regulations. These are also varying in di erent countries under di erent national regulations in trac management under low altitude, with 400 feet by FAA UTM and 700 feet in EASA U-space. Aerospace 2020, 7, x FOR PEER REVIEW 7 of 19 Aerospace 2020, 7, 140 7 of 20 3. Integrated Pilot Program for UAV Logistic Delivery 3. Integrated Pilot Program for UAV Logistic Delivery To examine the risk level of UAV flight operations, two cases of logistic delivery will be carried to look into the details of risk analysis. The IPP is a typical example for system performance To examine the risk level of UAV flight operations, two cases of logistic delivery will be carried to verification. look into the details of risk analysis. the IPP is a typical example for system performance verification. 3.1. UTM Environment for IPP 3.1. UTM Environment for IPP In the UTM system [10,29], the ADS-B like infrastructure plays an important role to cover 400 In the UTM system [10,29], the ADS-B like infrastructure plays an important role to cover 400 feet of regional UAS surveillance. The ADS-B like technology develops ground transceiver station feet of regional UAS surveillance. the ADS-B like technology develops ground transceiver station (GTS) and on-board unit (OBU) from 4G/LTE (long term evolution), APRS (automatic packet (GTS) and on-board unit (OBU) from 4G/LTE (long term evolution), APRS (automatic packet reporting reporting system), LoRa (long range wide area network), and XBee, for data link to UTM, as shown system), LoRa (long range wide area network), and XBee, for data link to UTM, as shown in Figure 2. in Figure 2. Airspace NAS ATM ADS-R UAS 978MHz National UTM >400 feet ANSP 1090MHz Pilot-in-the-loop CAA Regulations > 400 feet BVLOS to Integrated Airspace Pilot-in-command Flight Data < 400 feet VLOS + BVLOS Regional Airspace Regional Regulations Regional UTM Flight Data Internet ADS-B Like UAV UTM Cloud sUAS Flight Data UAV Flight Data <400 feet Pilot-in-command In-flight Communication via Mobile Phone Figure 2. The hierarchical unmanned aircraft system (UAS) trac management (UTM) in Taiwan [10,29]. Figure 2. The hierarchical unmanned aircraft system (UAS) traffic management (UTM) in Taiwan [10,29]. Similar to the 4G/LTE base transceiver station (BTS) in mobile communication, three other types of the ADS-B like technology need to deploy and build their specific GTSs for territorial coverage to Similar to the 4G/LTE base transceiver station (BTS) in mobile communication, three other types relay UAV surveillance data into the Internet to the UTM cloud. the first region UTM was constructed of the ADS-B like technology need to deploy and build their specific GTSs for territorial coverage to in Tainan City with five LoRa stations, as shown in Figure 3 [19] for proof of concept (POC). These five relay UAV surveillance data into the Internet to the UTM cloud. The first region UTM was GTS sites are CJCU, Yujing, Baihe, Yanshui, and Xigang, as marked in Figure 3. the flight tests have constructed in Tainan City with five LoRa stations, as shown in Figure 3 [19] for proof of concept been verified with full surveillance coverage in Tainan City under UTM operation [10,29]. (POC). These five GTS sites are CJCU, Yujing, Baihe, Yanshui, and Xigang, as marked in Figure 3. Two types of OBUs are selected for tests in this paper, a 4G/LTE cell phone and a LoRa OBU, as The flight tests have been verified with full surveillance coverage in Tainan City under UTM shown in Figure 4 [29], corresponding to its infrastructure. operation [10,29]. 3.2. Processes for IPP This study tries to examine the UAV flight trajectories to estimate risk prevention functions, such as risk assessment and UTM surveillance. In the operation, the integrated UAV risk prevention can be examined by Figure 5 in real time including air risk and ground risk [14,15,19]. the air risk is focused on UAV collision avoidance via DAA, while the ground risk concerns the human injury due to crash. For sUAS, time to conflict (TTC) among UAVs is conducted by the DAA software in UTM server [30]. the performance is feasible to implement since sUAS is not capable of carrying additional space or payload to carry the detection hardware on the airborne. APRS LTE LoRa X-Bee Aerospace 2020, 7, x FOR PEER REVIEW 8 of 19 Aerospace 2020, 7, 140 8 of 20 Aerospace 2020, 7, x FOR PEER REVIEW 8 of 19 Baihe Yanshui Baihe Yanshui Xigang Yujing Xigang Yujing CJCU CJCU Figure 3. Long range wide area network (LoRa) ground transceiver station (GTS) deployment in Tainan for the first proof of concept (POC). Figure Figure 3. 3. Long Long range range wide wide area area network network (LoRa) (Lo grRa ound ) grtransceiver ound transc station eiver st (GTS) ation deployment (GTS) deploym in Tainan ent in for Tainan for Tw the o first typ the f pr es oof of irs of OBUs t proof concept ar of co e (POC). se ncept lected (PO for C).test s in this paper, a 4G/LTE cell phone and a LoRa OBU, as shown in Figure 4 [29], corresponding to its infrastructure. Two types of OBUs are selected for tests in this paper, a 4G/LTE cell phone and a LoRa OBU, as shown in Figure 4 [29], corresponding to its infrastructure. (a) 4G (b) LoRa Figure Figure 4. Automatic 4. Automatic dependent depende surveillance–br nt surveillance oadcast –broadc (ADS-B) ast (ADS like -B) on-boar like on d -board units (OBUs) units (O for BU the s) for 4th the (a) 4G (b) LoRa generation mobile communi cation long term evolution (4G/LTE) and LoRa. 4th generation mobile communi cation long term evolution (4G/LTE) and LoRa. Figure 4. Automatic dependent surveillance–broadcast (ADS-B) like on-board units (OBUs) for the Referring to Allouch et al. [14] of a useful flight process, the UAS provider or operators should 3.2. 4th generation mobi Processes for IPP le communi cation long term evolution (4G/LTE) and LoRa. set up standard operation procedures (SOP) to carry pre-flight, in-flight and post-flight. In Taiwan, This study tries to examine the UAV flight trajectories to estimate risk prevention functions, such the UTM flight procedures are also with three similar phases. In the pre-flight phase, the CPD path 3.2. Processes for IPP as risk assessment and UTM surveillance. In the operation, the integrated UAV risk prevention can planning [11] and ELS [17] are applied to check referring to the territorial geodetic information to This study tries to examine the UAV flight trajectories to estimate risk prevention functions, such be examined by Figure 5 in real time including air risk and ground risk [14,15,19]. The air risk is by-pass the densely populated areas. the pre-flight procedure is shown in Figure 6. Pilots and their as risk assessment and UTM surveillance. In the operation, the integrated UAV risk prevention can focused on UAV collision avoidance via DAA, while the ground risk concerns the human injury due UAVs are required to log-in from the UAS management information system (MIS) by CAA [3]. the MIS be examined by Figure 5 in real time including air risk and ground risk [14,15,19]. The air risk is to crash. For sUAS, time to conflict (TTC) among UAVs is conducted by the DAA software in UTM includes databases of licensed pilots, registered UAVs, and no-flight zones (NFZ). In MIS, red, yellow, focused on UAV collision avoidance via DAA, while the ground risk concerns the human injury due server [30]. The performance is feasible to implement since sUAS is not capable of carrying additional and green areas are marked to identify restricted (red), conditioned (yellow), and free (green) airspace to crash. For sUAS, time to conflict (TTC) among UAVs is conducted by the DAA software in UTM space or payload to carry the detection hardware on the airborne. to fly. the pilots need to submit the flight plan into UTM for approval. UTM will check the proposed server [30]. The performance is feasible to implement since sUAS is not capable of carrying additional flight route to keep away from the red NFZ. Since multi-rotor UAVs can fly no longer than 60 min space or payload to carry the detection hardware on the airborne. at 8 m/s velocity at present, the flight route will not be farther than 28 km in its surveillance range. the communication test between controller to pilot should adopt either the 4G/LTE cell phone or Zello broadcast [10,29]. Aerospace 2020, 7, x FOR PEER REVIEW 9 of 19 Risk Assessment Corridor Path Planning Pre-Flight Geofensing Avoidance Expected Level of Safety Risk Prevention ADS-B Like Communication In-Flight UTM Surveillance Detect and Avoid Aerospace 2020, 7, 140 9 of 20 Aerospace 2020, 7, x FOR PEER REVIEW 9 of 19 Data Analysis Flight Performance Post-Flight Flight Operation Quality Risk Assessment Assurance Corridor Path Planning Pre-Flight Geofensing Avoidance Expected Level of Safety Figure 5. The concept of integrated UAV risk prevention system. Referring to Allouch et al. [14] of a useful flight process, the UAS provider or operators should set up standard operation procedures (SOP) to carry R pre isk -f li Pght, reve in nt-ifl oin ght and post-flight. In Taiwan, ADS-B Like Communication In-Flight the UTM flight procedures are also with three similar phases. In the pre-flight phase, the CPD path UTM Surveillance planning [11] and ELS [17] are applied to check referring to the territorial geodetic information to by- Detect and Avoid pass the densely populated areas. The pre-flight procedure is shown in Figure 6. Pilots and their UAVs are required to log-in from the UAS management information system (MIS) by CAA [3]. The MIS includes databases of licensed pilots, registered UAVs, and no-flight zones (NFZ). In MIS, red, Data Analysis Flight Performance yellow, and green areas are marked to identify restricted (red), conditioned (yellow), and free (green) Post-Flight Flight Operation Quality airspace to fly. The pilots need to submit the flight plan into UTM for approval. UTM will check the Assurance proposed flight route to keep away from the red NFZ. Since multi-rotor UAVs can fly no longer than 60 min at 8 m/s velocity at present, the flight route will not be farther than 28 km in its surveillance Figure 5. The concept of integrated UAV risk prevention system. Figure 5. The concept of integrated UAV risk prevention system. range. The communication test between controller to pilot should adopt either the 4G/LTE cell phone or Zello broadcast [10,29]. Referring to Allouch et al. [14] of a useful flight process, the UAS provider or operators should set up standard operation procedures (SOP) to carry pre-flight, in-flight and post-flight. In Taiwan, CAA UAV Management Information System the UTM flight procedures are also with three similar phases. In the pre-flight phase, the CPD path Registration Internet Pilot/Operator Sign-In planning [11] and ELS [17] are applied to check referring to the territorial geodetic information to by- Pilot/UAV Registered Pilot pass the densely populated areas. The pre-flight procedure is shown in Figure 6. Pilots and their Log-in UAV Databasee UAVs are required to log-in from the UAS management information system (MIS) by CAA [3]. The Log-in Registered MIS includes databases of licensed pilots, registered UAVs, and no-flight zones (NFZ). In MIS, red, Approval UAV Database yellow, and green areas are marked to identify restricted (red), conditioned (yellow), and free (green) Restrict Airspace Flight Plan airspace to fly. The pilots need to submit the flight plan into UTM for approval. UTM will check the Database Meteorological proposed flight route to keep away from the red NFZ. Since multi-rotor UAVs can fly no longer than Observation Flight Schedule 60 min at 8 m/s velocity at present, the flight route will not be farther than 28 km in its surveillance UTM Regional range. The communication test between controller to pilot should adopt either the 4G/LTE cell phone Center Flight Approval or Zello broadcast [10,29]. Payment E-Pay CAA UAV Management Information System ADS R -B eg L isitk re ation Controller-Pilot Internet Communication Communication Pilot/Operator Sign-In Initial Setup Initial Setup Pilot/UAV Test Call Registered Pilot Log-in Waypoint 0 Pilot Call iU nA V Databasee Verification Verification Communication Log-in Registered Cleared Approval UAV Database Approve to Take-off Via CPC Restrict Airspace Flight Plan Database Meteorological Figure 6. UTM pre-flight procedures. Observ Figure ation 6. UTM pre-flight procedures. Flight Schedule UTM Regional When the flight plan has been approved and the airspace is clear to go, the UTM controller will Center Flight Approval issue a clearance to the pilot for take-o , as shown in Figure 7. In the in-flight phase, UAVs have been equipped with an ADS-B like OBU for surveillance [10,29]. ADS-B like OBU broadcasts flight Payment E-Pay data down to GTS and connects into the UTM cloud. the UTM controller will monitor UAV flights ADS-B Like Controller-Pilot and o er flight tracking with no path violation. the DAA performance will be carried on the UTM Communication Communication Initial Setup Initial Setup server software with conflict detection and resolution. Once an UAV separation violates, the UTM Test Call Waypoint 0 Pilot Call in controller will intervene by the controller-pilot communication (CPC) for conflict resolution advisory Verification Verification Communication (RA) [10,30]. the UTM software DAA creates a mechanism similar to the separation bubble in TCAS, Cleared Approve to Take-off as shown in Figure 8. When multiple UAVs appear in a small window (range), the DAA will be Via CPC Figure 6. UTM pre-flight procedures. RA RA Aerospace 2020, 7, x FOR PEER REVIEW 10 of 19 Aerospace 2020, 7, x FOR PEER REVIEW 10 of 19 When the flight plan has been approved and the airspace is clear to go, the UTM controller will When the flight plan has been approved and the airspace is clear to go, the UTM controller will issue a clearance to the pilot for take-off, as shown in Figure 7. In the in-flight phase, UAVs have been issue a clearance to the pilot for take-off, as shown in Figure 7. In the in-flight phase, UAVs have been equipped with an ADS-B like OBU for surveillance [10,29]. ADS-B like OBU broadcasts flight data equipped with an ADS-B like OBU for surveillance [10,29]. ADS-B like OBU broadcasts flight data down to GTS and connects into the UTM cloud. The UTM controller will monitor UAV flights and down to GTS and connects into the UTM cloud. The UTM controller will monitor UAV flights and offer flight tracking with no path violation. The DAA performance will be carried on the UTM server offer flight tracking with no path violation. The DAA performance will be carried on the UTM server software with conflict detection and resolution. Once an UAV separation violates, the UTM controller software with conflict detection and resolution. Once an UAV separation violates, the UTM controller will intervene by the controller-pilot communication (CPC) for conflict resolution advisory (RA) Aerospace 2020, 7, 140 10 of 20 will intervene by the controller-pilot communication (CPC) for conflict resolution advisory (RA) [10,30]. The UTM software DAA creates a mechanism similar to the separation bubble in TCAS, as [10,30]. The UTM software DAA creates a mechanism similar to the separation bubble in TCAS, as shown in Figure 8. When multiple UAVs appear in a small window (range), the DAA will be shown in Figure 8. When multiple UAVs appear in a small window (range), the DAA will be activated. UAV speed extrapolations will check the possible time to conflict (TTC) for the next few data activated. UAV speed extrapolations will check the possible time to conflict (TTC) for the next few activated. UAV speed extrapolations will check the possible time to conflict (TTC) for the next few intervals. TTC generates warning signals for trac advisory (TA) and resolution advisory (RA), where data intervals. TTC generates warning signals for traffic advisory (TA) and resolution advisory (RA), data intervals. TTC generates warning signals for traffic advisory (TA) and resolution advisory (RA), TA = 48 s and RA = 25 s. In UTM, the surveillance data interval are regularly set from 8–10 s. Figure 8 where TA = 48 s and RA = 25 s. In UTM, the surveillance data interval are regularly set from 8–10 s. where TA = 48 s and RA = 25 s. In UTM, the surveillance data interval are regularly set from 8–10 s. shows the concept of DAA referring to TCAS by ICAO. DAA RA is performed by CPC. the priority Figure 8 shows the concept of DAA referring to TCAS by ICAO. DAA RA is performed by CPC. The Figure 8 shows the concept of DAA referring to TCAS by ICAO. DAA RA is performed by CPC. The assessment follows the air trac control rule. priority assessment follows the air traffic control rule. priority assessment follows the air traffic control rule. UAVs UAVs ADS-B Like UAV Take-off ADS-B Like UAV Take-off Clearance Surveillance Data Clearance Surveillance Data UTM UTM Center UTM UTM Center Controllers Main Server Controllers Main Server Google Map Google Map Internet Surveillance Internet Surveillance UTM Cloud UTM Cloud Situation Situation Awareness Awareness Air Navigation Contact Pilot Air Navigation DAA/Detour Contact Pilot Service DAA/Detour Service Waypoint Waypoint Management Management Internet Internet Landing CAA UAV Landing CAA UAV Report Specification Report Specification Management Management System Mission Complete Auto Log Book System Mission Complete Auto Log Book Log-Out Log-Out Figure 7. UTM in-flight and post-flight procedures. Figure 7. UTM in-flight and post-flight procedures. Figure 7. UTM in-flight and post-flight procedures. TA TA CPA CPA Intruder Intruder Heading/Speed GPS x, y, z Heading/Speed GPS x, y, z TTC TTC TTC= 60 sec Surveillance Circle TTC= 60 sec Surveillance Circle about 600 m radius about 600 m radius Figure 8. Concept of UTM detect and avoid (DAA) mechanism. Figure 8. Concept of UTM detect and avoid (DAA) mechanism. Figure 8. Concept of UTM detect and avoid (DAA) mechanism. The UTM server extrapolates UAV headings to check their separation by TTC. If a conflict is The UTM server extrapolates UAV headings to check their separation by TTC. If a conflict is The UTM server extrapolates UAV headings to check their separation by TTC. If a conflict is possible in the next few time intervals, an alert will generate to the UTM controller. The UTM possible in the next few time intervals, an alert will generate to the UTM controller. The UTM possible in the next few time intervals, an alert will generate to the UTM controller. the UTM controller controller will check the priority and contact the less priority pilot to detour via CPC. CPC is activated controller will check the priority and contact the less priority pilot to detour via CPC. CPC is activated will check the priority and contact the less priority pilot to detour via CPC. CPC is activated using using a cell phone or Zello broadcasting. In Figure 9, AK 1035 has less priority to detour by a right using a cell phone or Zello broadcasting. In Figure 9, AK 1035 has less priority to detour by a right a cell phone or Zello broadcasting. In Figure 9, AK 1035 has less priority to detour by a right turn of turn of 15 degrees to avoid. turn of 15 degrees to avoid. 15 degrees to avoid. In the post-flight phase, the UAV pilots need to log out from the UTM and CAA MIS. the ADS-B like reports 90 byte data including pilot ID, UAV ID, GPS position, and six-DoF flight data. It appears as follows: [Heading(5); UAV(6); Pilot(6); Lat.(9); Long.(10); Alt.(4); 6 DoF(p, q, r, , , ,) (36); V(6); A(6); Tail(2)]. the short one is only UAV ID, Pilot ID, and X, Y, Z data. After flights, the flight performance can be analyzed for flight operation quality assurance (FOQA) using the UTM surveillance data. CPC CP E Cm E em rg ee rn gc ey n c C ya C llall UU AA S S S e Sr ev riv cie c e Aerospace 2020, 7, 140 11 of 20 Aerospace 2020, 7, x FOR PEER REVIEW 11 of 19 MX1122, 50 MX1122, 50 TTC<60 sec AK1035, 50 AK1035, 50 DAA Resolution DAA Alert TTC>100 sec (a) (b) Figure 9. Conflict alert and resolution, less priority UAV avoids (a) TTC < 60sec with DAA alert; (b) Figure 9. Conflict alert and resolution, less priority UAV avoids (a) TTC < 60sec with DAA alert; TTC > 100 s with DAA resolution. (b) TTC > 100 s with DAA resolution. 3.3. UAV Delivery Case Study In the post-flight phase, the UAV pilots need to log out from the UTM and CAA MIS. The ADS- B like reports 90 byte data including pilot ID, UAV ID, GPS position, and six-DoF flight data. It This study concerns the use of UAV for logistic delivery in case studies. the flight system and test appears as follows: specifications are described in detail. the purpose of the delivery case tries to examine the e ectiveness [Heading(5); UAV(6); Pilot(6); Lat.(9); Long.(10); Alt.(4); 6 DoF(p, q, r, α, β, γ,) (36); V(6); A(6); of surveillance capability of ADS-B like infrastructure under UTM, and further to analyze UAV expected Tail(2)]. The short one is only UAV ID, Pilot ID, and X, Y, Z data. After flights, the flight performance levels of safety (ELS). Two cases are demonstrated for the logistic delivery scenario. can be analyzed for flight operation quality assurance (FOQA) using the UTM surveillance data. An hexa-rotor UAV is used to deliver a parcel with a weight of 2.4 kg (four bottles of water). the UAV flies at a velocity of 5–8 m/s and cruises at 35–50 m AGL for less than 40 min. the flight 3.3. UAV Delivery Case Study performance adopts a Pixhawk flight control autopilot with a Google map mission planner for beyond visual line of sight (BVLOS). the flight is fully monitored by UTM using ADS-B like technology. This study concerns the use of UAV for logistic delivery in case studies. The flight system and The first flight case is flying in Ping Tung County. UAV carries the 4G/LTE cell phone for flight test specifications are described in detail. The purpose of the delivery case tries to examine the operation and video surveillance. a 900 MHz communication is added for the control uplink. a QR code effectiveness of surveillance capability of ADS-B like infrastructure under UTM, and further to is placed on the ground as the final target. the video downlink aims at the target with image processing analyze UAV expected levels of safety (ELS). Two cases are demonstrated for the logistic delivery in the QR code recognition to accurately locate the delivery target via 4G/LTE. This scenario performs scenario. the UAS delivery over a river, where ground transportation detours a long router. the surveillance An hexa-rotor UAV is used to deliver a parcel with a weight of 2.4 kg (four bottles of water). The uses 4G/LTE to report the flight track via the Internet to UTM. UAV flies at a velocity of 5–8 m/s and cruises at 35–50 m AGL for less than 40 min. The flight In the second flight case, the UAV delivery is from CJCU to Hsin-Ta Harbor. the delivery goes performance adopts a Pixhawk flight control autopilot with a Google map mission planner for directly from CJCU to the destination. However, ground transportation needs to detour with several beyond visual line of sight (BVLOS). The flight is fully monitored by UTM using ADS-B like junctions. the flight is monitored under the Tainan RUTM. the UAV flight data are collected and technology. broadcast into ground transceiver station (GTS) via the Internet to the UTM cloud. This test uses The first flight case is flying in Ping Tung County. UAV carries the 4G/LTE cell phone for flight the LoRa ADS-B like on-board unit (OBU) for real time cloud surveillance [10,19]. operation and video surveillance. A 900 MHz communication is added for the control uplink. A QR code is placed on the ground as the final target. The video downlink aims at the target with image 3.4. Risk Mitigation from Path Planning processing in the QR code recognition to accurately locate the delivery target via 4G/LTE. This scenario performs the UAS delivery over a river, where ground transportation detours a long router. In the pre-flight phase in Figure 6, the first flight case was operated in a remote area, where The surveillance uses 4G/LTE to report the flight track via the Internet to UTM. the density of population results in less safety concerns. In the first scenario, Figure 10 shows delivery In the second flight case, the UAV delivery is from CJCU to Hsin-Ta Harbor. The delivery goes across a river in Santiman, Pingtung County. the UAV delivers a small parcel of 2.4 kg to the ultra-light directly from CJCU to the destination. However, ground transportation needs to detour with several field at the other side of river, which is about 1.6 km away. Since the area is a countryside, the route is junctions. The flight is monitored under the Tainan RUTM. The UAV flight data are collected and planned point to point and is flown at a velocity of 5–8 m/s. the test was conducted on a clear sunny broadcast into ground transceiver station (GTS) via the Internet to the UTM cloud. This test uses the day, wind miles/h from azimuth 320. the UAV flew a fair wind to cross the river. the actual flight time LoRa ADS-B like on-board unit (OBU) for real time cloud surveillance [10,19]. was about 8 min and the UAV flew at a constant altitude of 35 m above ground level (AGL). the 8-min UAV flight would require 25 min for ground transportation. In this flight test, 4G/LTE of the selected 3.4. Risk Mitigation from Path Planning ADS-B like is adopted for surveillance into the UTM. This test just tries to verify 4G/LTE as a choice of ADS-B like technology wherever the BTS can cover. Figure 10a shows the path planning and 5b shows In the pre-flight phase in Figure 6, the first flight case was operated in a remote area, where the the real flight surveillance. density of population results in less safety concerns. In the first scenario, Figure 10 shows delivery across a river in Santiman, Pingtung County. The UAV delivers a small parcel of 2.4 kg to the ultra- light field at the other side of river, which is about 1.6 km away. Since the area is a countryside, the route is planned point to point and is flown at a velocity of 5–8 m/s. The test was conducted on a clear sunny day, wind miles/h from azimuth 320. The UAV flew a fair wind to cross the river. The actual Aer Aer osp osp ace ace 20 20 20 20 , , 77 , , x x FO FO R P R P EE EE R R RE RE VIEW VIEW 12 12 of of 19 19 fl fl ight ight time time was was about about 8 8 m m in in and and th th e e U U AV AV ff lew lew at at a a con con stant stant al al ti ti tude tude of of 35 35 m m above above groun groun d d leve leve l l (AGL). The 8-min UAV flight would require 25 min for ground transportation. In this flight test, (AGL). The 8-min UAV flight would require 25 min for ground transportation. In this flight test, 4G/LT 4G/LT E E of of th th e e sel sel ee ct ct ed ed A A DS DS -- B B li lke ike is is ad ad op op ted ted for for ss urvei urvei llance llance in in to to th th e e UTM. UTM. Th Th is is test test ju ju st st tries tries to to verify 4G/LTE as a choice of ADS-B like technology wherever the BTS can cover. Figure 10a shows verify 4G/LTE as a choice of ADS-B like technology wherever the BTS can cover. Figure 10a shows Aerospace 2020, 7, 140 12 of 20 th th e path e path pl pl an an ning ning and 5b and 5b s s how how s the re s the re al f al f lig lig ht ht s s urve urve illa illa nce nce . . Pa P ra ar ca hc u h tu e te Landing Landing TT ak ae k-e o -fo ff f Flight Path Flight Path Po Pio n itnt Pla P n la nn in ng in 1g .6 1 5. 6 k 5m km Path Plan Path Plan Ultralight Ultralight Airfield Airfield FliF g lh ig t h T t rT ac ra kck Delivery Delivery Point Point (( aa )) (( b b )) Figure 10. UAV delivery test 1 at Santiman across the river under 4G to UTM (a) Test 1 path Figure 10. UAV delivery test 1 at Santiman across the river under 4G to UTM (a) Test 1 path Figure 10. UAV delivery test 1 at Santiman across the river under 4G to UTM (a) Test 1 path planning; pla pla nnin nnin g; ( g; ( b b ) T ) T est 1 est 1 UTM t UTM t ra ra ck ck ing ing . . (b) Test 1 UTM tracking. In the second scenario, the delivery takes place in a suburban area at Queiren, Tainan. The UAV In In th the e second second scenar scenario, io, th the e deli delivery very ta takes kes pl place ace in ina assuburban uburban ar ar eea a at at Queiren Queiren, , T T ainan. ainan.Th the e UA UAV V deli deli deliver ve ve red red ed a a a par p p ar ar cel ce ce l l fr ff rom om rom CJCU CJ CJ CU CU (Chang (Ch (Ch ang ang Jung JJ ung ung Christian Ch Ch rr is is ti ti an University) an Univer Univer si si ty to ty )) Hsin-T to to Hs Hs in a in - Harbor - T T a a Ha Ha ,rbor, rbor, which whic whic hash h ahas dir has ect a a direct distance of about 8 km, as shown in Figure 11. Figure 11a shows the path planning with GTS direct distance dis of tanc about e of 8abo km, ut as 8 shown km, as in shown Figur in e 11 Figure . Figur 11. e 11 Fi agshows ure 11a the shpath ows th planning e path pla with nnin GTS g w coverage ith GTS cov cov under era era ge ge Tainan und und er er RUTM. T T ai ai n n an an the R R U U TM. test TM. uses Th Th ee test LoRa test uses uses OBU Lo Lo R to R a a r O O eport BU BU to to UA repo repo V surveillance rt rt UAV UAV sur sur ve ve data ill ill ance ance to LoRa d d ata ata to to GTS. LoR LoR In aa GTS. In Figure 11a, the CJCU GTS has its coverage of 15 km. The arrows show the range within the GT Figur S. e In11 Fig a,u the re 11a CJCU , the GTS CJCU has GT its S ha coverage s its cov of erag 15ekm. of 15 the km. arr Th ows e arshow rows show the range the ra within nge wit the hin GTS the GT GT coverage. S S cov cov era era ge T ge est . Test 2 . Test 2 2 was wa wa flown s fl s fl own own on on a on a a clear cle cle ar sunny d sunny ar sunny d dayay ay with with with wind wind 3 wind 3 3 m m at a m at a at azimuth zim zim uth uth 020. 0 0 20. It is 20. It is It is a a a ffair f air wind air wind wind for flight. A geo-fence mechanism was activated to keep the UAV away from the restricted area of for for flight. flight. A a geo-fence geo-fence mechanism mechanism was was activated activated to to k keep eep th the e UA UAV V away away fr from om the the r restr estricted icted ar are ea a o of f th th the e e ai airport. ai rpo rpo rt. rt. T T This his his test test test f f li flies li ee s s ov ov over er er th th the e e freew freew freeway ay ay . .. On On Onth th the e e de de delivery li li very very p p path ath ath pl pl planning, anni anni ng ng , , th th the e e U U UA AV AV Vh has h as as an an an al altitude al titude titude trajectory constraint to pass the freeway by 50 m above. The total flight time was 24.5 min, cruising traject trajectory ory con constraint straint to to p pass ass th the e fr freew eeway ay by by 50 50 m m abo above. ve. Th the e total total flight flight time time was was 24.5 24.5 min, min, cr cruising uising at 50 m AGL. For test 2, the UAV took off at 35 m elevation and landed at 5 m elevation at Hsin-Ta at at 50 50 m m AGL. AGL. For For test test 2, 2, th the e U UA AV V took took o o ff at at 35 35 m m elevation elevation and and landed landed at at 5 5 m m elevation elevation at at H Hsin-T sin-Ta a Harbor. In this flight scenario, it takes 67 min on the ground traffic. Since the second scenario is Ha Harbor rbor. . In In th this is flight flight scen scenario, ario, it it ta takes kes 67 67 min min on on the the gr groun ound d tra tra ffic. c. S Since ince th the e second second scenario scenario is is performed near the airport, the UTM controller will monitor the UAV logistic operation from performed performed near near the the airport, airport, theth UTM e UTM contr coller ontroll will er monitor will mothe nitor UA tV he logistic UAV operation logistic op frer om atio intr n usion from intrusion into the restrict area. intrus into the ion restrict into th ar e ea. restrict area. Airport Restricted Area Airport Restricted Area CJCU Take-off CJCU Take-off Chang Jung Chang Jung ChC rih sr tiia st n ia U nn U iv n .iv. UT U M T M GT G ST S at CJCU at CJCU 10 1 k0 m km Flight Path Flight Path 10 km PlP al nan nin nig n g 8 8 k m km 10 km Hsin-Ta Hsin-Ta HT Harbor HT Harbor HaH rb ao rb r or Landing 10 Landing 5km 5km (( aa )) (( b b )) Figure Figure Figure11 11 11. . .UAV UAV UAV de de delivery lili very very t es ttest es t t2 2 2 from from from CJC CJC CJCU, U U , ,Queiren Queiren Queiren to to to Hsi Hsi Hsin-T nn -Ta -Ta Harbor aHarbor Harbor un un under de de r r UTM UTM UTM ( ( aa ) () aTest )Test Test 2 2 2p p path ath ath planning; (b) Test 2 UTM tracking. planning; (b) Test 2 UTM tracking. planning; (b) Test 2 UTM tracking. 4. Risk Assessment for UAS Delivery From the logistic delivery tests, the selected territories are either remote or suburban with less population. This is the most feasible condition for UAS logistic delivery at present. How is the risk level to adopt UAV into logistic delivery? Using FAA simulations, this paper examines the expected level of safety (ELS) in these two demonstrations. Flight Track Flight Track at la P h P n Path Plan Aerospace 2020, 7, 140 13 of 20 4.1. FAA Risk Level The risk assessment is given in terms of the FAAs report in 2015. the risk level is evaluated by the probability of events. a sUAV (<25 kg) failed to free-fall to the ground [1]. the FAA scenario applies to the UAV flying above a certain population density (n/m ) concerning the MTBF for the specific UAV, the frontal impact area of the UAV (S ), the impact area of humans (S ), the kinetic energy of UAV h the sUAV (KE), the exposed fraction (EF) of humans, and the probability of lethality (P ) for impact casualties for an UAV with MTOW (M) and flight velocity (V). the population density is calculated by the total ground area of the surface (S ) to the number of humans (n). The KE of the sUAV is determined using the terminal velocity of the sUAV, the MTOW (M), and the drag coecient (C = 0.3), as: KE = MV (1) The FAAs risk level is calculated using the probability of a sUAV event as: S ( ) EF P UAS l P = (2) event MTBF where: Population Density = (3) In the pre-flight phase of Figure 6, this paper uses the assessment formula and modifies the assumptions for the real situation to calculate the risk level for the delivery scenario. Two flight test results at Santiman, Pingtung and Queiren, Tainan were used to estimate the safety level of delivery using the population density figures for 2018 for suburban and remote areas from the Ministry of Interior (MOI), Taiwan [31]. the risk assessment data are listed as follows: a. Selected UAV: Arm pitch 83 cm hexa-rotor, MTOW M = 12 kg, cruise speed V = 8 m/s. b. Population: n/S = 0.00013 (Santiman/remote area, 130 n/km ) and 0.000651 (Queiren/suburban area, 651 n/km ). c. MTBF: 100 h, in accordance with the FAA data [1,4]. d. Area of UAV (S ): 0.6889 m , arm pitch 0.83 0.83 m from a hexa-rotor. UAV e. Exposed Fraction (EF): 0.2, in accordance with the FAA data [1,4]. f. Probability of lethality (P ): 0.3, in accordance with the FAA data [1,4]. g. Kinetic energy of UAV (KE): 384 Joules, from a hexa-rotor. h. MTOW (M): 12 kg. i. Velocity (V): 8 m/s (maximum operating speed of hexa-rotor H83). The results are shown in Table 1. the most significant di erences are seen in the KE, M, and V terms. Referring to the FAA simulation uses 250 g at a speed of 25 m/s in an area with a high population 2 2 (10,000 n/m or 3853 n/km ), the flight tests in this paper use the hexa-rotor H83 flying with 12 kg at 8 m/s speed and at an altitude of 35 m. In terms of the population density, the probability of an event in a remote area (Santiman, PT) is less than that for a suburban area (Queiren, TN) [31]. In comparison, the respective risk level for a commercial air transport and general transport is 1 10 E/h and 5 8 7 5 10 E/h, and the risk levels of this study, 5.37 10 and 2.69 10 , in Table 1 are reasonable. According to the FAAs statistical data for general aviation (GA) fatal accident rates from 2010 to 2017 [12], the average GA fatal accident rate is 1.028 10 per 100,000 flight hours. the probability for an UAV that is calculated by this study is less than the GA fatal accident rate [12]. the lack of safety data for UAV has no direct e ect on the reliability and safety level compared with the GA and commercial air transport safety records. Experience shows that due to the gyroscopic e ect of the multi-rotor system [11], the quad-rotor does fall in a spiral trajectory. the impact to the ground is di erent from that of free-fall. Table 1 shows that the terminal velocity of the sUAS is 25.7 m/s [1,11]. Aerospace 2020, 7, x FOR PEER REVIEW 14 of 19 Aerospace 2020, 7, 140 14 of 20 lack of safety data for UAV has no direct effect on the reliability and safety level compared with the GA and commercial air transport safety records. Experience shows that due to the gyroscopic effect This is an exaggerated result in terms of the real performance for a sUAS. This study uses a maximum of the multi-rotor system [11], the quad-rotor does fall in a spiral trajectory. The impact to the ground terminal velocity of 8 m/s [11] from the actual flight experience. is different from that of free-fall. Table 1 shows that the terminal velocity of the sUAS is 25.7 m/s The UAVs weight and speed (KE), the population density, and the failure rate have the greatest [1,11]. This is an exaggerated result in terms of the real performance for a sUAS. This study uses a impact on the number of casualties [13]. the results of the risk assessment for this study show that maximum terminal velocity of 8 m/s [11] from the actual flight experience. future research should focus on the reasons for the failure of an UAV. UAV performance parameters, such as MTOW and velocity, determine the severity of an UAV failure for a specific population density, Table 1. Comparison of risk level for this study and using FAA data [1]. so the reliability of an UAS must be improved in terms of the MTBF. Further study of risk prevention Population MTBF Suav Pl KE M V and risk management for UAS services is necessary to increase the safety to the public. Test Area UAV Pevent EF 2 2 Density(n/m ) (h) (m ) (%) (J) (kg) (m/s) −8 Urban, FAA sUAS 4.68 × 10 0.0039 100 0.02 0.2 0.3 82.47 0.25 25.7 4.2. Analysis on Expected Level of Safety (ELS) −8 Santiman, PT Hexa-rotor 5.37 × 10 0.00013 100 0.6889 0.2 0.3 384 12 8 The level of safety is calculated for the purpose of integrating di erent types of UAV into −7 Queiren, TN Hexa-rotor 2.69 × 10 0.000651 100 0.6889 0.2 0.3 384 12 8 the NAS regarding the safety requirements. Based on Weibel’s ground impact model [21] in Figure 12, the possible injury and fatality are taken into account. Moreover, the expected level of safety (ELS) The UAVs weight and speed (KE), the population density, and the failure rate have the greatest is calculated using Equation (4). In Figure 12, the air risk model is also descriptive to the failure of impact on the number of casualties [13]. The results of the risk assessment for this study show that DAA. Two UAVs might collide to crash to turn into this ground impact model. It is also important that future research should focus on the reasons for the failure of an UAV. UAV performance parameters, the UTM controller is highly responsible to pay attention to the DAA alert when multiple UAVs appear such as MTOW and velocity, determine the severity of an UAV failure for a specific population in the same area. In Figure 8, the UAV icon on the UTM display can extrapolate by their heading density, so the reliability of an UAS must be improved in terms of the MTBF. Further study of risk arrows by 5 to 10 times of the surveillance data period of 5–8 s. It will examine the TTC of two UAVs prevention and risk management for UAS services is necessary to increase the safety to the public. by their TAs and following RAs. 4.2. Analysis on Expected Level of Safety (ELS) ELS = A P (1 P ) (4) exp p pen mit The level of safety is calculated for the purpose of integrating different types of UAV into the MTBF NAS regarding the safety requirements. Based on Weibel’s ground impact model [21] in Figure 12, ELS: Expected level of safety (failure/hour). the possible injury and fatality are taken into account. Moreover, the expected level of safety (ELS) is MTBF: Mean time between failure of UAV. calculated using Equation (4). In Figure 12, the air risk model is also descriptive to the failure of DAA. A : Area of exposure (m ), or frontal impact area (FA) of UAV. Tw exp o UAVs might collide to crash to turn into this ground impact model. It is also important that the : Population density. UTM p controller is highly responsible to pay attention to the DAA alert when multiple UAVs appear P in th : e Pr sobability ame area. of In penetration, Figure 8, th re otor UAV UAiV con of 0.25 on th [13 e U ,24 TM ]. display can extrapolate by their heading pen arrows by 5 to 10 times of the surveillance data period of 5–8 s. It will examine the TTC of two UAVs P : Probability of mitigation preventing ground fatality, rotor UAV of 0.75 [24]. mit by their TAs and following RAs. Harm to Failure of Debris Resulting Impact in Public on UAV Penetration of Penetration Populated Area? Ground System? sheltering? Fatal? Recovery None No Accident None No Exposure to Debris None No Penetration Possible Injury Yes Fatality Figure 12. Ground impact model [21]. Figure 12. Ground impact model [21]. Based on Equation (4), the ELS of outcome in this study is shown as Table 2. 𝐸𝐿𝑆 = 𝐴 𝜌 𝑃 (1 − 𝑃 ) (4) 𝑒𝑥𝑝 𝑝 𝑝𝑒𝑛 𝑚𝑖𝑡 𝐵𝑀𝑇𝐹 In Table 2, the ELS is relatively high in areas with large populations. However, according to the frontal impact area of di erent UAVs [11], the smaller the UAV, the lower the risk. In this study, ELS: Expected level of safety (failure/hour). the unsafety MTBF: Mean rank of time ELS betw for T een able failu 2 ar re e of Queir UAen, V. Santiman, and FAA area. Aexp: Area of exposure (m ), or frontal impact area (FA) of UAV. ρp: Population density. Ppen: Probability of penetration, rotor UAV of 0.25 [13,24]. Aerospace 2020, 7, 140 15 of 20 Table 1. Comparison of risk level for this study and using FAA data [1]. Population MTBF S P KE M V uav l Test Area UAV P EF event 2 2 Density (n/m ) (h) (m ) (%) (J) (kg) (m/s) Urban, FAA sUAS 4.68 10 0.0039 100 0.02 0.2 0.3 82.47 0.25 25.7 Santiman, PT Hexa-rotor 5.37 10 0.00013 100 0.6889 0.2 0.3 384 12 8 Queiren, TN Hexa-rotor 2.69 10 0.000651 100 0.6889 0.2 0.3 384 12 8 Table 2. Comparison of the expected level of safety (ELS) for this study with the ground impact model. Population 1/MTBF S uav P (1 P ) Test Area UAV pen ELS mit 2 2 Density (n/m ) (h) (m ) Urban, FAA sUAS 0.0039 0.01 0.02 0.25 0.25 4.875 10 Santiman, PT Hexa-rotor 0.00013 0.01 0.69 0.25 0.25 5.597 10 Queiren, TN Hexa-rotor 0.000651 0.01 0.69 0.25 0.25 2.803 10 Aerospace 2020, 7, 140 16 of 20 4.3. Risk Prevention Tool for the UTM System The UAV flight operation must be real time monitoring in the in-flight phase, and analysis in the post-flight phase via the UTM system, as shown in Figure 5. For this reason, this study selects the ADS-B like communication infrastructure in the developing UTM by introducing 4G/LTE and LoRa for flight test 1 and 2, respectively, for sUAs [10,29]. During the in-flight phase, UTM provides the UAV surveillance function using ADS-B like OBU. Unlike manned aircraft, the sUAS are not detectable via radar or other independent surveillance techniques. ADS-B like surveillance on UTM shall be feasible to develop [17]. Mobile communication is the most a ordable communication system to adopt, for its wide area deployment. the 4G/LTE cell phone is most available to adopt into UAVs either using cell phones or modules. However, 4G/LTE will not guarantee to cover as high as 400 feet. the developing hierarchical UTM [10] system with the ADS-B like communication adopts devices [21] in high reliability, light weight, low cost, and wide coverage through gateway deployment. Other than 4G/LTE, the LoRa and other proposed technology require constructing and GTS deployment to relay radio surveillance from UAVs into the UTM cloud [10,30,31]. the proposed ADS-B like GTSs receive all UAV surveillance data into the UTM cloud, and distributes to the regional UTM (RUTM) for local governments. In the UTM operation, DAA is another key function to build. Using software manipulation, DAA can e ectively intervene to the controller-pilot communication (CPC) for pilot manipulation to avoidance [30]. With big data collection, UTM FOQA would definitely o er a great contribution to UAV/UAS risk assessment in the future. In the UTM, the real time communication delay by a few seconds, as shown in Table 3, from flight experiences using ADS-B like surveillance infrastructure should be paid attention to. These data will be key factors to a ect the UAS risk assessment. Table 3. Real time delay in UTM. ADS-B Like Period Tx/Rx Cloud UTM C-P 4G/LTE 6~8 0.8 0.8~1 1~2 6~10 LoRa 6~10 1~2 1~2 1~2 6~10 APRS 5~13 4~8 2~4 1~2 6~10 Xbee 6~10 1~2 1~2 1~2 6~10 in seconds 5. Conclusions This paper demonstrates an UAV risk assessment using the case study on logistic delivery in remote and suburban areas. Referring to the FAA TCL certification [27], the population density is a major concern to approve the UAS into flight services. the case studies release UAVs in the BVLOS flight. By way of careful path planning, the UAV services can be feasible to meet TCL 3 to reach the accepted risk level to the public. In this paper, the UAS flight surveillance is e ectively monitored under UTM using the ADS-B like infrastructure operating BVLOS under UTM surveillance. Google routing is used for path planning to keep away from the highly populated areas and NFZ. a restricted area and geo-fence is also created to fulfill the UTM requirement for UAVs flying below 400 feet. the ADS-B like infrastructure using 4G/LTE or LoRa is adopted for surveillance with excellent performance. In terms of the risk impact of UAV services to the public, the risk assessment calculates the risk level for a UAV logistic delivery service. the results accomplish very high confidence in the UAV logistic delivery flights with e ective data surveillance on UTM. The risk assessment uses the FAA task force’s recommendation [1] for simulations. In terms 8 7 of actual UAV operations and flight scenarios, the risk level is higher (10 –10 E/h) than that in the manned air transportation system. However, this result is still acceptable for UAV delivery Aerospace 2020, 7, 140 17 of 20 application in remote areas and in suburban areas. the real operation of an UAV parcel delivery service in di erent areas is verified to be ecient and feasible for safe operation. From the flight tests, the UTM system can completely monitor the logistic delivery flights. It is confident for risk assessment with e ective and transparent flight surveillance. the demonstrations use 4G/LTE and LoRa for UTM surveillance. the ADS-B like infrastructure would be more dependent and reliable in surveillance coverage [10,29,30]. Referring to the GTS deployment in Tainan City of Figure 3, the developing ADS-B like infrastructure can o er a seamless UTM surveillance, which mitigates the risk of failure during flight operations. The safety level assessment using the database for manned aircraft is not realistic because the velocity of a VTOL sUAV is set at 25.7 m/s [1]. For low altitude flights, the feasible velocity lies within 8 m/s for the multiple rotor UAVs. In conclusion, the risk assessment for UAS using a case study of logistic deliveries in remote and suburban areas demonstrates an acceptable measure for the expected level of safety (ELS). It is feasible to use the UAS routine services for logistic delivery services in Taiwan. This paper merits flight experiments to calculate the expected level of safety using UAVs for logistic delivery under the UTM environment. the result supports the applications of UAV logistic delivery in Taiwan. Funding: This work is financially supported by the Ministry of Science and Technology under contract no. MOST109-2622-E-309-001-CC1. Acknowledgments: The flight tests are conducted by the UTM team in CJCU by Chin E. Lin to support this paper. Conflicts of Interest: There is no conflict of interest to any institutes or individuals, since this is an academic research program. Nomenclature a Acceleration including Fall Drag Air density at sea level (kg/m ) H Cruise Height (m) V Cruise Velocity, V = V V Cruise Velocity or Initial Velocity (m/s) C Drag coecient g Gravitational acceleration (m/s ) M Maximum Take-o Weight (MTOW) or Mass (kg) y Initial Altitude (m), y = H 0 0 V Initial velocity (m/s) V Terminal velocity (m/s) FIA Frontal Impact Area (m ) 2 2 KE Kinetic Energy (kg m /s ) Abbreviations 3D Dirty, Danger, and Dull 4G/LTE 4th Generation Mobile Communication Long Term Evolution AC Advisory Circular ADS-B Like Automatic Dependent Surveillance–Broadcast Like AGL Above Ground Level APRS Automatic Packet Reporting System ARC Air Risk Classifications ATC/ATM Air Trac Control/Air Trac Management BVLOS Beyond Visual Line of Sight CJCU Chang Jung Christian University CPC Controller-Pilot Communication CPD Closest Point Detection Aerospace 2020, 7, 140 18 of 20 DAA/SAA Detect (Sense) and Avoid DoF Degrees-of-Freedom ELOS Equivalent Level of Safety ELS Expected Level of Safety ESC Electronic Speed Control (Converter) FAA Federal Aviation Administration, USA FHA Functional Hazard Assessment FSM Function Safety Management GA General Aviation GTS Ground Transceiver Station HTOL Horizontal Take-o and Landing (Fixed Wings) ICAO International Civil Aviation Organization IPP Integrated Pilot Program ISO International Standard Organization JAA Joint Aviation Authority, Europe JARUS Joint Authorities for Rulemaking of Unmanned System KE Kinetic Energy LiPo Lithium Polymer Battery LoRa Long Range Wide Area Network MAC Mid-Air Collision MEMS Microelectromechanical Sensors MRO Maintenance, Repair and Overhaul MTBF Mean Time between Failures MTOW Maximum Take-o Weight NAS National Airspace System NIA Non-Integrated Airspace OBU On-Board Unit RA Resolution Advisory RUTM/NUTM Regional UTM/National UTM SORA Specific Operation Risk Assessment TA Trac Advisory TCAS Trac Alert and Collision Avoidance System TCL Technical Capability Level TLS Target Level of Safety TTC Time to Conflict UAV Unmanned Aerial Vehicle UAS Unmanned Aircraft System UTM UAS Trac Management VTOL Vertical Take-o and Landing (Rotor Wings) References 1. FAA. Unmanned Aircraft System (UAS) Registration Task Force (RTF), Aviation Rulemaking Committee (ARC), Task Force Recommendations Final Repot. Available online: https://www.faa.gov/uas/resources/ event_archive/media/micro-uas-arc-final-report.pdf (accessed on 20 November 2015). 2. FAA. UAS Regulation; FAA Website: Washington, DC, USA, 2016. 3. Civil Aeronautical Administration (CAA). Aviation Regulations in UAV Special Chapter. April 2018. Available online: https://www.caa.gov.tw/article.aspx?a=188&lang=1 (accessed on 28 February 2020). 4. FAA. Small Unmanned Aircraft Systems (sUAS); AC 107-2; FAA: Washington, DC, USA, 2016. 5. Kopardekar, P.; Rios, J.; Prevot, T.; Johnson, M.; Jung, J.; Robinson, J.E. Unmanned Aircraft System Trac Management (UTM) Concept of Operation; AIAA Aviation Technology, Integration, and Operations (ATIO): Washington, DC, USA, 2016. Aerospace 2020, 7, 140 19 of 20 6. Amazon Prime Air. Amazon Revising the Airspace Model for the Safe Integration of small Unmanned Aircraft Systems. Available online: https://utm.arc.nasa.gov/docs/Amazon_Revising%20the%20Airspace% 20Model%20for%20the%20Safe%20Integration%20of%20sUAS[6].pdf (accessed on 3 July 2019). 7. Business Insider Intelligence. Amazon Gives Prime Air Public Demo, 27 March 2017. Available online: http://www.businessinsider.com/amazon-gives-prime-air-public-demo-2017-3 (accessed on 3 May 2019). 8. EASA. Civil Drones (Unmanned Aircraft). Available online: https://www.easa.europa.eu/easa-and-you/civil- drones-rpas (accessed on 18 December 2019). 9. SESAR Joint Undertaking. Blueprint U-Space, SESAR Joint Undertaking; Publications oce of the European Union: Luxembourg, 2017. 10. Lin, C.E.; Shao, P.-C.; Lin, Y.-Y. System Operation of Regional UTM in Taiwan. Aerospace 2020, 7, 65. [CrossRef] 11. Lin, C.E.; Shao, P.-C. Failure Analysis for an Unmanned Aerial Vehicle Using Safe Path Planning. J. Aerosp. Inf. Syst. 2020, 17, 358–369. [CrossRef] 12. FAA. Fact Sheet–General Aviation Safety. Available online: https://www.faa.gov/news/fact_sheets/news_ story.cfm?newsid=21274 (accessed on 30 July 2018). 13. Melnyk, R.; Schrage, D.; Volovoi, V.; Jimenez, H. a third-party casualty risk model for unmanned aircraft system operations. Reliab. Eng. Syst. Saf. 2014, 124, 105–116. [CrossRef] 14. Allouch, A.; Koubaa, A.; Khalgui, M.; Abbes, T. Qualitative and Quantitative Risk Analysis and Safety Assessment of Unmanned Aerial Vehicles Missions Over the Internet. IEEE Access 2019, 7, 53392–53410. [CrossRef] 15. Joint Authorities for Rulemaking of Unmanned System (JARUS). JARUS Guidelines on Specific Operations Risk Assessment, 2nd ed.; JARUS: Cologne, Germany, 2019. 16. Melnyk, R.; Schrage, D.; Volovoi, V.; Jimencz, H. Sens and AVOID Requirements for Unmanned Aircraft Systems Using a Target Level of Safety Approach. Risk Anal. 2014, 34, 1894–1906. [CrossRef] [PubMed] 17. Weibel, R.E.; Hansman, R.J. Safety considerations for operation of di erent class of UAVs in the NAS. In Proceedings of the AIAA 3rd Unmanned Unlimited Technical Conference, Chicago, IL, USA, 20–23 September 2004. 18. Dalamagkidis, K.; Valavanis, K.P.; Piegl, L.A. Evaluating the risk of unmanned aircraft ground impacts. In Proceedings of the 2008 16th Mediterranean Conference on Control and Automation; Institute of Electrical and Electronics Engineers, Ajaccio, France, 25–27 June 2008; pp. 709–716. 19. Barr, L.C.; Newman, R.; Ancel, E.; Belcastro, C.M.; Foster, J.V.; Evans, J.; Klyde, D.H. Preliminary Preliminary Risk Assessment for Small Unmanned Aircraft Systems. In Proceedings of the 17th AIAA Aviation Technology, Integration, and Operations Conference, Denver, CO, USA, 5 June 2017. [CrossRef] 20. JAA/EUROCONTROL. a Concept for European Regulations for Civil Unmanned Aerial Vehicles, Appendix 3.5. 2004. Available online: https://www.easa.europa.eu/sites/default/files/dfu/NPA_16_2005_Appendix.pdf (accessed on 24 July 2019). 21. Dalamagkidis, K.; Valavanis, K.P.; Piegl, L.A. On Integrating Unmanned Aircraft Systems into the National Airspace System, 2nd ed.; Springer: Berlin/Heidelberg, Germany, 2012. 22. Ancel, E.; Capristan, F.M.; Foster, J.V.; Condotta, R.C. Real-Time Risk Assessment Framework for Unmanned Aircraft System (UAS) Trac Management (UTM). In Proceedings of the 17th AIAA Aviation Technology, Integration, and Operations Conference, Denver, CO, USA, 5 June 2017; American Institute of Aeronautics and Astronautics (AIAA): Washington, DC, USA, 2017. 23. Lum, C.; Waggoner, B. A Risk Based Paradigm and Model for Unmanned Aerial Systems in the National Airspace; Infotech@Aerospace: St. Louis, MI, USA, 2011. 24. Grimsley, F.M. Equivalent Safety using Casualty Expectation Approach. In Proceedings of the AIAA 3rd Unmanned Unlimited Technical Conference, Chicago, IL, USA, 20–23 September 2004. 25. Garrett-Glaster, B. EASA Propose Risk-Based Airworthiness Standards for Light Unmanned Aircraft. Aviation Today, 21 July 2020. Available online: https://www.aviationtoday.com/2020/07/21/easa-proposes-risk-based- airworthiness-standards-light-unmanned-aircraft/ (accessed on 21 July 2020). 26. Aweiss, A.; Owens, B.; Rios, J.; Homola, J.; Mohlenbrink, C. Unmanned Aircraft Systems (UAS) Trac Management (UTM) National Campaign II. In Proceedings of the AIAA-2018-1727, 2018 AIAA Science and Technology Forum (SciTech), Kissimmee, FL, USA, 8–12 January 2018. 27. Schrage, D. Integration of Functional Safety Management and Systems Engineering for Development Assurance. Engineering 2017, 1–33. Aerospace 2020, 7, 140 20 of 20 28. Martin, T.; Huang, Z.F.; McFadyen, A. Airspace Risk Management for UAVs a Framework for Optimising Detector Performance Standards and Airspace Trac using JARUS SORA. In Proceedings of the 2018 IEEE/AIAA 37th Digital Avionics Systems Conference (DASC), London, UK, 23–27 September 2018; pp. 1516–1525. 29. Lin, Y.-H.; Lin, C.E.; Chen, H.-C. ADS-B Like UTM Surveillance Using APRS Infrastructure. Aerospace 2020, 7, 100. [CrossRef] 30. Lin, C.E.; Shao, P.-C. Development of Hierarchical UAS Trac Management (UTM) in Taiwan. J. Phys. Conf. Ser. 2020, 1509, 012012. [CrossRef] 31. Pingtung City Government. Pingtung City Government: Pingtung City Population Density. Available online: http://www.ptcg.gov.tw/Default.aspx (accessed on 5 June 2019). © 2020 by the author. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).
Journal
Aerospace – Multidisciplinary Digital Publishing Institute
Published: Sep 25, 2020