123doku.com

J Intell Robot Syst (2021) 103: 16decisions in a detailed audit trail to make it repeatable andallow improvements where the methodology lacks rigor andvalidity. For some steps, we had to take pragmatic decisionsin order to move forward. These decisions are documented,and the steps can be revisited with more detailed researchand more rigor.Our method to find, select, and evaluate sources forSections 3 and 5 leans on Systematic Literature Reviews(SLRs) [6]. As academic sources, we used the onlinedatabase IEEE Xplore5 and open access articles fromvarious publishers found in the databases of arXiv6 andGoogle Scholar. Furthermore, we used articles found andarchived during previous, UAS-related works of ours.Occasionally we searched for non-scientific publications onGoogle7 to enhance our assessments. All search queriesused for this study are presented in the Appendix A. A foundpublication had to satisfy all of the following inclusioncriteria to be selected for further examination:–––––The publication addresses UAS in its title or abstractusing any relevant term or description.It discusses at least one UAS application in the contextof public safety.It contains explicit information about this application’smaturity. For example, it describes performed testflights, demonstrations, or successful real-life missions.It must contain at least a descriptive concept ofthe application and not only preliminary technologyresearch.It was published in 2010 or later. We suspected morerecent publications are more likely to exhibit highermaturity levels.As Fig. 2 illustrated, the number of publicationsconcerned with UAS in the public safety context is vast andseems to increase steadily. If—despite our best efforts tocreate sufficiently targeted queries—an academic databasesearch returned more than 100 results, we sorted the listby relevance (how closely the results matched the query, asdetermined by the search engine itself) and used only thefirst 100 results for further processing. This procedure wasespecially required for Google Scholar, which a) presentedmore than 10,000 results for a majority of the queries and b)does not provide an option for sorting other than relevance.To save time in the later steps, we preselected the resultsobtained here by their titles.3 Meta-review of UAS in Public SafetyApplicationsPrevious studies reviewed public safety applications ofUAS. Next to the inclusion criteria listed in Section 2,5 https://ieeexplore.ieee.org/6 https://arxiv.org7 https://www.google.comPage 3 of 15 16found studies had to meet the criterion of actually being(or at least containing) a review on UAS applications. Thiscriterion was added because the search term “survey” alsoyielded articles using other meanings of the word (as in“land survey”).3.1 Comprehensive ReviewsUAVs as Mobile Sensing Platforms (MSPs) in the domainssmart cities & public safety as well as civil security &disaster response were reviewed in [7]. The survey listsnumerous case studies for various applications, leadingto the conclusion that “such mobile sensing / actuationplatforms have already matured to the point where theyare widely considered as a viable addition to existingapplications and approaches” [7]. The group stated thatthe emergence of UAV swarms is inevitable, and the“usefulness of such swarms will dramatically increase withgrowing autonomy”. This is similar to conclusions foundin a review about optimization approaches for civil UAVapplications: “future research should work out dynamicplanning schemes” for a range of relevant drone operations[and] develop approaches to deal with data uncertainty [8].In [9] the author presented a survey on proposed andimplemented drone applications in Africa, where manylocal governments actively support the technology. Examples include delivering healthcare-related goods, wildlifemonitoring, and even predicting atrocities in conflict-riddenregions. The study lists the benefits and challenges ofdrones before naming considerations for proper and safedrone usage, including community participation, consistentregulatory frameworks, and more.3.2 UAS for Disaster ManagementSeveral reviews specifically focused on the disasterresponse domain. Reference [10] lists case studies ofUAVs used for imagery collection following hurricanes,typhoons, and earthquakes. The authors concluded thatUAVs are viable data collection tools for such events.Reference [11] presented a systematic literature reviewof UAVs in humanitarian relief applications, charting thisresearch field’s rapid growth through a statistical analysis.Out of 117 surveyed papers, 62 addressed the recoverystage of “natural sudden-onset” disasters. Furthermore,they showed that most surveyed publications focused onimproving the equipment’s performance, especially forimagery and mapping applications. A similar study onthe use of UAS in humanitarian relief was conductedin [12]. The authors summarized that while the use ofUAS is more widespread in the military, humanitarian aidprojects are mostly still in the concept and testing phase.UAS developments and concepts in this area take place

16 Page 4 of 15in the use cases mapping (e.g., of flood risks), deliveries(e.g., medical products), search and rescue (e.g., afternatural catastrophes), monitoring changes (e.g., buildingsdestroyed after earthquakes), public health prevention(e.g., combating mosquitoes plagues), agriculture (e.g.,monitoring crops), monitoring climate change (e.g., monitorglaciers with thermal cameras), demining (e.g., searchingfor shifted mines after floods) and protecting civilians andpeacekeeping (e.g., detect hidden troops) [12].State of the art in UAV-based photogrammetry8 (especially for earthquakes, volcanic activity, and landslides) wasstudied in [13]. The authors concluded that this techniquehas proven useful in various scenarios and will continue togain traction. Reference [14] examined the combination ofUAVs and Wireless Sensor Networks (WSNs) for disastermanagement, highlighting open issues in network stability. The authors analyzed 29 case studies in terms of theaddressed disaster stage, used technology (WSN or UAV),and performed application (monitoring, information fusion,situational awareness, damage assessment, standalone communication system, search and rescue) [14, Tab. 1]. Worthmentioning is furthermore [15, Sec. 2.G], which brieflylists UAV usage examples from disasters such as the 2011Fukushima Daiichi nuclear disaster or the 2013 earthquakein Port au Prince, Haiti.J Intell Robot Syst (2021) 103: 16showed that agencies are making attempts to add UASexpertise and skilled drone pilots” [19].A sociological study on institutional realities and publicperceptions of police use of UAVs in Canada was presentedin [20]. The authors examined service flight logs andconcluded that most operational flights were “dedicatedto tasks associated with assisting persons in immediaterisk (e.g., missing persons) or gathering evidence for anidentified crime” [20].Multiple countries employed UAS in attempts to combatthe COVID-19 pandemic. A collection of related use cases,exemplifying how versatile UAS can be, was presented in[21]: crowd surveillance, public announcements, temperature screenings, spraying disinfectants, and the delivery ofmedical supplies.3.5 Contribution of this StudyThe listed reviews provide insight into the state of theart in various applications of UAS from a dominantlytechnological point of view. In contrast, the review at handaims to develop a more in-depth analysis of the maturitylevels, or “market-readiness”, of UAS-based public safetyapplications. To the best of our knowledge, no previousstudy has attempted a comparable investigation in thisdomain.3.3 UAS for Forest Fire MonitoringThree of the selected reviews focused on forest firemonitoring. The authors of [16] reviewed the developmentof UAV-based forest firefighting systems and presentan overview of vision-based fire detection techniques.Reference [17] states that this application is an active fieldof research with numerous experiments already performed,but with plenty of room for improvement before it canbe considered sufficiently mature—especially regarding firedetection algorithms, integration, and testing under realisticconditions. The authors of [18] emphasized increasedautonomy as a key enabler for this use case.3.4 Other ReviewsA recent survey on drone usage in public safety agencieswas performed by members of the U.S. Department of Commerce’s National Institute of Standards and Technology[19]. In line with the research group’s focus, it stressed theneed for improved wireless communication during missionsand gathered requirements for future, UAV-based solutions.Additionally, the survey “identified a potential gap in theuse of drones in some public safety operations where dronesmight improve [.] operational performance, but it also8 Generatingmeasurements or maps from photographs.4 Application Readiness LevelsComparing the maturity levels of various applicationsrequires a consistent metric. For this purpose, we decidedto employ the nine-step Application Readiness Level (ARL)index conceived by NASA to track and manage a project’sprogression [22, 23]. ARLs are an adaptation of thewell-known Technology Readiness Levels (TRLs) [24] forassessing technology development and risk. The mainreason we chose ARLs over TRLs was to emphasize ourfocus on the maturity of applications, which can depend ona variety of technologies, rather than the UAS technology onits own. Another small advantage of ARLs for this study isthe more explicit consideration of the end user’s perspective,e.g., including the need for documentation and training.However, ARLs and TRLs are sufficiently similar to eachother that they can be used almost interchangeably. In thefollowing, we briefly recite the ARL concept to improvethe readability of the later sections. Figure 3 presents anoverview.ARLs can be divided into three major phases:Phase I (ARL 1–3):Phase II (ARL 4–6):Discovery and feasibility.Development, testing, and validation.

J Intell Robot Syst (2021) 103: 16Page 5 of 15 16ARL 8:ARL 9:The application, in its final form, is provento work under expected conditions. The userdocumentation is complete.This application system has seen successful andsustained use in the operational environment.Generally, ARLs are not meant to measure the performanceof individual tools or models—they are used to quantifytheir readiness for use in a decision-making process. Therestill can be significant room for improvement after reachingARL 9, e.g., increased performance or lower costs.5 Application Categories and their MaturityLevelsFig. 3 Overview of Application Readiness Levels. Adapted from[22, p. 2] to fit this article’s layoutPhase III (ARL 7–9):Integration into partner’s system.Many research projects end before or in the developmentphase (II), whereas the final integration and fine-tuning ofthe system (phase III) is often left to companies or the endusers. The individual levels can be summarized as follows:ARL 1:ARL 2:ARL 3:ARL 4:ARL 5:ARL 6:ARL 7:Basic scientific concepts and insights areobserved and reported, providing the foundationfor application ideas.The formulation of the application concept andthe individual components has begun. The fullsystem is still speculative.Feasibility studies are conducted, providing aproof-of-concept. The components have beentested and validated independently.Basic components are integrated into a prototypesystem to verify that they will work together.The prototype is integrated with reasonably realistic supporting elements. The potential advantages of the new application are clearly articulated.The prototype and its advantages have beendemonstrated in a relevant or simulated operational environment.The prototype and all pre-deployed componentsare fully integrated, tested, and demonstrated inthe end-user’s operational environment to win thepartner’s confidence.As previously stated in Section 2, the vast number of UASapplications renders individual examinations infeasible.However, many applications show substantial similarities,allowing one to sort them into a manageable list ofapplication categories. For example, tasks to transportmaterial are mostly similar from a UAV’s perspectiveregardless of this material being vaccines, humanitarian aid,fire extinguishing agents, or something else entirely. Severalcategorizations have been proposed or re-utilized in theliterature [7, 11, 14, 20]. Although good starting points,we found these categorizations inadequate for our reviewbecause they include areas outside of the public safetydomain (e.g., precision agriculture) or miss applicationcategories that we deem relevant (e.g., intercepting roguedrones). We have therefore derived a novel categorization,primarily based on the respectively required capabilities andhardware components. The categories are namely RemoteSensing, Mapping, Monitoring, Human-drone Interaction(HDI), Flying Ad-Hoc Networks (FANETs), Transportationand Counter UAV Systems (C-UAS). An overview ofthese categories, including exemplary technical and missiongoals, is presented in Table 1. A proper validation ispending, but we have confidence in this proposal since ita) overlaps with the existing categorizations and b) couldsuccessfully be applied to all references we found for thesubsequent maturity assessments.Each of the following subsections describes one application category and presents this category’s maturity assessment. The assessments themselves arose from the followingprocess:1. We performed a literature search (see Section 2) forpublications addressing applications of the category9 .Naturally, we included findings from the existingreviews (Section 3).9 Theexact queries and other details are listed in the Appendix A.

16 Page 6 of 15J Intell Robot Syst (2021) 103: 16Table 1 Overview of UAS application categories used in this studyApplication categoryTechnical goalExemplary mission goalExamples of requiredhardware and capabilitiesRemote SensingGather image or othersensor data.Reconstruct a digitalmap of an area.Continually monitor anarea over a long-termperiod.Gain an aerial overview,find missing persons.Assess structural damages.Cameras or other sensors.MappingMonitoringDetect developing forest fires.Human-drone InteractionCommunicate with people in the UAV’s vicinity.Guide civilians towardsevacuation routes.Flying Ad-Hoc NetworksCreate a radio networkbetween UAS and userson the ground.Deliver cargo.Provide cellphone service to civilians in need.Stop rogue drones fromflying.Ensurethesafetyof starting or landing crewed aircraft atairports.TransportationCounter UAV Systems2. In each article, we investigated which ARL characteristics (Section 4) were fulfilled. For example, anarticle describing a successful demonstration in an operational environment was said to reach ARL 6. A definiteassignment to an ARL was omitted if it became evident that previously assessed publications in the samecategory scored higher.3. The maturity of a category is the maximum ARLreached by one of the publications selected for it.Further investigations for the category were omitted if apublication reached the highest possible ARL 9.Not necessarily all the articles selected by us forparticular categories are referenced, but only those whichdirectly influenced our assessments.Limitations Our ARL ratings of the proposed applicationcategories (and the individual publications within) arepotentially biased and should be treated with due caution.Despite our best efforts to achieve scientific rigor andobjectivity, a certain degree of uncertainty and bias cannotbe ruled out for a number of reasons, e.g.:––Our assessments are limited by the literature wediscovered.The authors of investigated publications (especially inthe context of commercial or preliminary applications)Provide humanitarian aidto people in hard-toaccess areas.–Same as Remote Sensingplus data fusion.Same as Remote Sensing plus increased autonomy, automatic recharging, pattern recognition.Human-MachineInterfaces, e.g., displays,speakers,ormicrophones.Antennas, routers, automated recharging.Attachablecontainer,possibly pick-up andrelease or dispersionmechanism.Radar, net cannons, signal jammers.may have overstated positive and understated (or evenomitted) negative results.The assessments are of qualitative nature and as suchalso prone to personal biases.Nonetheless, we find the overall picture emerging fromthe entirety of all assessments to be coherent and haveconfidence in the conclusions we have drawn from it(Sections 6 and 7).105.1 Remote SensingDescription Many UAS applications are purely concernedwith data acquisition and require the UAVs to carrynot much more than one or more sensors. We groupthese applications into the Remote Sensing category, alsoidentified by the term Imagery (if no sensors other thancameras are required). Variants of this category includeSituational Awareness (Fig. 4), Searching (e.g., for missingpersons), and Inspection (e.g., to assess damages). In ouropinion, these variants are mostly discerned by the chosenflight paths and required level of detail in the data but areotherwise equivalent.10 Wewould like to invite inclined readers to discuss the assessments,and point to relevant publications we missed, by writing an email tothe authors.

J Intell Robot Syst (2021) 103: 16Page 7 of 15 16Therefore, we have furthermore decided to split Mappinginto two subcategories.5.2.1 Assessment MappingDescription On one side of the spectrum, there arescenarios where a map of a medium-sized area (e.g., a singlebuilding and its immediate surroundings) is desired, withplenty of time available for data acquisition and processing.Applications fitting this description fall into the AssessmentMapping subcategory, which can also be seen as a mapbased expansion of Inspection.Fig. 4 Drone view of a major fire in Dortmund overlayed with infraredmeasurements. The drone was used to track the fire’s progression.Found via [28], photo by courtesy of the Dortmund Fire DepartmentAssessment This category is arguably the most mature, asnumerous public safety authorities routinely use UAS forthis purpose nowadays. Supplementary to examples foundin existing reviews [10, 15, 20, 21], we would like tohighlight the following news articles:–––Tracking the progression of flames and finding optimalpositions for fire hoses during the 2019 fire of NotreDame de Paris [25].Searching for missing persons [26].Assessing a criminal suspect’s weaponry [27].The Remote Sensing or Imagery category has maturedto a point where periodicals for professionals such as theGerman Feuerwehr-Magazin [Fire Department Magazine]have published special issues about UAS for firefighters andsimilar professions, including overviews of commerciallyavailable models and legal considerations [28]. Thesustained use clearly warrants a 9 on the ARL scale.5.2 MappingIf the remotely gathered data is processed and fused tocreate a digital model of the environment, we assign theapplication to the Mapping category. The created maps mayrepresent the physical structure of the environment but couldalso model other types of data, such as temperatures orconcentrations of harmful substances [29]. Mapping canbe seen as an extension of Remote Sensing—there is oftenno distinction made in the literature. Nevertheless, due tothe increased requirements on processing capabilities andthe high number of publications explicitly addressing thistopic, we decided that a designated category is justified.The difficulty of Mapping applications strongly depends onthe available mission time and environmental complexity.Assessment Numerous articles have shown the viability ofUAV-based Assessment Mapping in a variety of scenarios,such as geological disasters (earthquakes, volcanic activity,and landslides) [13, 15, 30, 31], nuclear disasters [15], firerisk estimation [32], traffic accident reconstruction [33],or crime scene investigation [34]. Most publications inthis area focus on post-disaster assessment [11], but casestudies addressing the mitigation and preparation stagesexist as well [32, 35]. A recent review on establishedand emerging technologies for structural damage mappingshowed the substantial evolution this field has seen inthe last decades [36]. The quality of maps produced inthe aftermath of earthquakes can allow civil engineers toestimate the risk of building collapse [31]. Most often, theUAVs are flown manually or along pre-defined paths [36].Nowadays, there are multiple professional photogrammetrysoftware solutions readily available for this purpose [37].Consequential to the sustained use, we determined theoverall ARL for Assessment Mapping applications to be 9.5.2.2 Emergency MappingDescription On the other side, a requirements workshop weconducted with members of the Institute of Fire Serviceand Rescue Technology Dortmund pointed to the desirefor mapping capabilities during the response stage oflarge-scale disasters, such as fires in industrial complexesor flash floods [38]. This application area is referredto as Emergency Mapping, the “creation of maps, geoinformation products and spatial analyses dedicated toproviding situational awareness emergency managementand immediate crisis information for response” [39] orsometimes Real-Time Mapping [40]. Such scenarios cannecessitate the usage of a swarm of UAVs utilizing adistributed system architecture [41, 42] since a singlevehicle system might not be able to cover the area fast ordetailed enough. At the same time, the number of availablehuman pilots is typically limited. Hence, the UAS wouldhave to operate autonomously. I.e., the UAVs need tonavigate in complex terrain (e.g., with numerous physical

16 Page 8 of 15J Intell Robot Syst (2021) 103: 16of targeted scenarios. While high maturity levels have beenreached for 2D mapping from an unobstructed bird’s-eyeview (ARL 7 for [43]), 3D mapping applications in moredemanding environments have just begun to move intophase II (development, testing

[email protected] Carsten Wolff [email protected] Christof Rohrig [email protected] 1 Institute for the Digital Transformation of Application and Living Domains (IDiAL), Dortmund University of Applied Sciences and Arts, Otto-Hahn-Str. 23, 44227 Dortmund, Germany 2 Faculty of Information Technology, Dortmund .Author: Merlin Stampa, Andreas Sutorma, Uwe Jahn, Jörg Thiem, Carsten Wolff, Christof RöhrigPublish Year: 2021