Abstract
Wilderness Search and Rescue (WSAR) focuses on locating and extricating missing persons in remote settings. As unmanned aerial vehicle (UAV) or “drone” technology has evolved, so has the literature describing its application in WSAR operations. We conducted a scoping review of literature that describes the use of UAVs in WSAR contexts. The Joanna Briggs Institute Framework for scoping reviews was followed using the Preferred Reporting Items for Systematic Reviews and Meta-Analyses Extension for Scoping Reviews method. Additional individual databases, article reference lists, and relevant grey literature were also included in the search to provide an impartial scope. Seven hundred forty-seven articles were identified. Of these, 56 were found to be duplicates. The remaining 691 were further screened and checked for eligibility. Ultimately, 21 studies were found that met our inclusion criteria. This literature supports the use of UAVs to increase the safety and efficiency of a WSAR operation for locating victims, assessing risks, carrying equipment, and restoring communication systems. Unmanned aerial vehicles are a potentially useful adjunct in the management of WSAR operations. Their limitations include objects obscuring victims, weather changes, uneven terrain, battery-limited flight time, and susceptibility to environmental damage.
Keywords
Introduction
Wilderness Search and Rescue (WSAR) teams are often deployed to search for, locate, and extricate individuals who are overdue, missing, lost, and/or possibly injured in remote areas and wilderness settings. 1 Wilderness Search and Rescue operations are generally considered to be time-critical, for delays in locating the victim may decrease their probability of survival. 1 Consequently, the search and locate phase of a wilderness rescue operation commonly requires the urgent sourcing and allocation of manpower and related resources, some of which may not be immediately available.1,2
Over the past few years, unmanned aerial vehicles (UAVs) or “drones” have shown the potential to shorten search and locate times, accelerate rates of intervention, minimize risks for rescuers and victims, and present a cost-effective alternative to conventional WSAR search techniques. 2 -4 Proponents of the application of UAV technologies in WSAR contexts cite a number of potential advantages for the rescue team.2,3,5 This said, there are also a number of limitations to their use. Consequently, decisions to purchase and operate UAVs in resource-constrained rescue settings should be defended against scientific evidence of their value and role in such contexts. We conducted a scoping review of the literature that describes the use of UAVs in WSAR contexts. This scoping review provides an assessment of the available evidence relating to the potential role and value of using UAVs in WSAR operations, including some of their limitations.
Methods
We followed the methodology and steps of the Joanna Briggs Institute Framework for the conducting of scoping reviews. 6 The Preferred Reporting Items for Systematic Reviews and Meta-Analyses Extension for Scoping Reviews (PRISMA-ScR) flow diagram was used to guide the data collection processes. 7 The approaches and steps followed in the context of our study are summarized below.
Step 1: Defining the Title
The title we derived to guide our review became “Use of Unmanned Aerial Vehicles for Wilderness Search and Rescue Operations: A Scoping Review.”
Step 2: Developing the Review Question
The research question we developed for this study became “What literature is available that describes the use of UAVs in WSAR operations?”
Step 3: Deciding on the Format and Approach
Our results are described according to the PRISMA-ScR flow diagram. 7
Step 4: Defining the Inclusion and Exclusion Criteria
Our inclusion criteria became any literature focusing on UAVs in WSAR operations. We excluded literature on UAVs outside of WSAR contexts and literature that was not available in English.
Step 5: Defining the Search Strategy
The following databases were used: Cumulated Index to Nursing and Allied Health Literature (CINAHL), Cochrane, Medline, and Scopus. The reference lists of the identified articles were further screened for additional literature, and any relevant grey literature that was found was also included. Initially, the following combination of terms and keywords was used: “drones,” “unmanned aircraft vehicles,” “unmanned aerial vehicle,” “remotely piloted vehicle,” “unmanned aviation vehicles,” and “wilderness search and rescue.” During the searches, additional terms and keywords were discovered. Where relevant, these were also considered to broaden the initial search. The search was conducted between July 2022 and August 2022.
Step 6: Data Selection and Management
The PRISMA-ScR selection process flow diagram for scoping reviews was used to guide and illustrate the data selection process of this review. 7 After identifying data from the databases, the literature and relevant grey literature were imported to an Excel spreadsheet, where all duplicates were removed. The titles of the literature were then screened; those titles that fit the brief together with nonspecific titles were retained for abstract screening. After title and abstract screening, the full-text versions of the remaining manuscripts were critically reviewed for inclusion considering the aim and objectives of the study.
Step 7: Data Extraction
Data extraction was completed using extraction forms and tables. The Joanna Briggs Institute data extraction form 6 was used to guide the researcher in identifying data sources, characteristics, and results. Each of these was tabulated under the headings included in Table 1.
Summary of the data extracted from the included records
UAV, unmanned aerial vehicle; USA, United States of America; WSAR, Wilderness Search and Rescue; NO, Norway; CH, Switzerland; PO, Poland; IT, Italy; SAR, search and rescue; TR, Turkey; UK, United Kingdom; AT, Austria; AOS, airborne optical sectioning; GR, Greece; SA, South Africa; KR, Korea.
Same victim.
Step 8: Analysis of the Evidence
The data were extracted, synthesized, mapped, and interpreted to determine frequency of characteristics, concepts, and occurrences.
Step 9: Presentation of the Results
Our results are presented in the form of tables, charts, and narrative summaries in the sections that follow.
Results
The results of the scoping review are summarized in Table 1. Of the 276 articles identified through database searches of Scopus (187), PubMed (82), CINAHL (5), and Cochrane (2), 21 records were found that met the inclusion criteria.2,3,5,8-25 The data extracted from these 21 records are summarized in Table 1. The final 3 columns of Table 1 summarize the different authors’ pronouncements on the type, usefulness, terrain, and context of the UAVs featured in their study.
Date of Publications and Country of Origin
Thirteen records were published within the past 5 y,5,14-25 7 records in the last 6 to 10 y,2,3,9-13 and 1 record was found that was older than 10 y. 8 No records were excluded based on their age. The majority of records included in the review came from the United States of America, contributing 7 records,2,5,8,13,16,19,20 and Austria with 3 records.17,21,22 Turkey,3,25 Italy,12,22 and Poland 11,15 contributed 2 records each. One record had overlapping origins, namely, Austria and Italy. 22 There was 1 record contribution each from Norway, 9 England, 14 Switzerland, 10 Greece, 18 Korea, 24 and South Africa. 23
Core Focus and Aims
The records identified revealed similar characteristics with regard to their focus and aim. Seven records focused on searching for a real victim with the support of a UAV.2,5,13-16,23 There were 4 records that focused on the general feasibility of utilizing a UAV in WSAR operations.8,9,20,22 Two records compared the usage of UAVs with classic ground SAR techniques.3,25 The development and testing of a system or software that could be used to improve the usability of a UAV in WSAR was the focus of 8 records. 10 -12,17-19,21,24
Methodology
The studies referred to in the records could be divided into retrospective (n=7) and prospective designs (n=14). The designs were further divided by method and approach; these included case reports (n=7) and experimental studies (n=14). The experimental studies further comprised feasibility studies (n=8), usability studies (n=3), and randomized simulation studies (n=3). The mean sample size for all the records was 29 simulated operations and patients and 1 patient in authentic contexts. Six records focused on simulated operations, and 7 studies reported on real-life operations. Eight records did not specify a sample size, nor did they include human participants.
Environment and Context
All the records identified reported on operations that were conducted in wilderness settings (as per inclusion and exclusion criteria). However, as the term “wilderness environment” was not clearly defined in the literature in terms of terrain, it became difficult to accurately link the different studies to a specific wilderness environment based on the information that was provided in the records.
Types of UAVs Used
Various models of UAVs were described throughout the records, each having different specifications. The majority of the UAVs featured in the records could be considered “entry” to “mid-level” battery-powered units. This is in contrast to larger “high-end” commercial and military type UAVs, which have bigger payloads, extended flight times, and are considerably more expensive. The UAVs conventionally used in WSAR contexts have payloads and operation times that are seen as “limited” in comparison to high-end units, with the majority of authors highlighting their flight times as a significant limitation. A summary of these is presented in Table 1.
Discussion
Wilderness Search and Rescue teams are commonly required to locate and rescue individuals who are lost, ill, or injured in remote wilderness environments.1,8 Wilderness settings differ considerably by climate, vegetation, and terrain. The type of terrain and environment has a great impact on the WSAR operations. Often, it is the harsh environment itself that creates the need for rescue in the first place. 8 The use of UAVs may play a role in limiting the unnecessary exposure of rescuers to environmental risks and potentially shorten the time taken to conduct a WSAR operation.
Role of UAVs in Risk Mitigation
Identifying and limiting risk is important in all rescue operations. Karaca et al, 3 Weldon and Hupy, 19 and Abrahamsen 9 found that utilizing UAVs in WSAR operations can improve risk identification and mitigation. Unmanned aerial vehicles can be deployed to survey and search areas that are not immediately accessible or too dangerous to rapidly clear by ground personnel. Scouting and analyzing areas in advance will also provide rescuers with better situational awareness, allowing risk mitigation strategies to be implemented for rescuers who are required to physically access and search the area.4,9,19
Having a good communication system is crucial in any WSAR mission.20,22 An inability to communicate with team members effectively poses several risks. The feasibility of using a UAV to restore radio communication during SAR operations was investigated by McRae et al. 20 The study conducted multiple field tests where a UAV equipped with a radio repeater was used to restore radio communication between SAR personnel. 20
One record describes a rescue in the Himalayas where a UAV was successfully used to locate a missing climber, ultimately leading to his rescue.5,13-15 The authors indicated that use of the UAV removed the need for human searchers to search a large area that was potentially unstable, unnecessarily.5,13-15 Similarly, Nomalanga 23 agreed that UAVs serve a valuable function in risk mitigation strategies by reporting on the use of the South African Western Cape Government Department of Health’s UAV in different wilderness rescue incidents, including a search for a victim who had jumped off a cliff into the ocean. Although the victim was never located, the use of the UAV meant that the risks associated with launching and use of boats and rescue swimmers in a dangerous area were avoided. 23 Van Tilburg 2 described how in Oregon a UAV was used to confirm a fatality in a canyon (using images acquired by the pilot), thus eliminating the need for rescuers to urgently climb down into the canyon. Rather, they were able to plan and access the area the following day.
Reducing the Time Taken to Locate the Victim
Wilderness Search and Rescue operations are seen to be time-critical as the victims’ survival probability decreases with time.3,12 We found the literature to be replete with examples of how UAVs have allegedly sped up the process of locating the victim. Hanrahan 16 reported on how the police department of Enfield, Connecticut, searched for a blind man who had been missing for more than 30 h in winter. The use of a UAV allowed the rescue teams to locate the man within half an hour.
Karaca et al 3 and Cicek et al 25 compared the differences between human searchers and UAVs in WSAR. Their studies used field experiments to compare search times between classic human search techniques and UAV searches separately. Although these studies used simulated rescue incidents, the results support the findings of Nomalanga 23 and Hanrahan 16 that UAVs allow victims to be located faster.
Impact of Advances in UAV Technology
Unmanned aerial vehicles are becoming ever more upgraded and equipped with additional accessories that can widen their scope and improve their usability and detection rates in WSAR operations. 3 Rodriguez 18 and Cacace et al 10 explored the use of different multirobot systems to control and navigate multiple UAVs simultaneously. The feasibility of the software was tested to autonomously allocate UAVs to individual tasks, eliminating the need for multiple UAV pilots and rescuers to search areas.9,18
Six studies mentioned utilizing UAVs equipped with thermal imaging.9,11,17,21,22,24 Of these, the core aim of 2 of the studies was to explore the feasibility of thermal imaging to detect victims and to confirm if they were still alive.11,24 In WSAR contexts, thermal imaging was found to be a useful aid in detecting live victims. 17 Although these technologies may be helpful, manual scanning of aerial images (in hopes of detecting a victim) was found to be challenging and time-consuming. Manual scanning has been shown to come with a risk of “false detection” due to human error and developing mental fatigue. 19 The development and use of identification software to assist with detecting victims was found to be helpful. 19 Identification software makes use of algorithms to detect victims. Five records were found that focused on human detection algorithms, and the authors all seem to come to a similar conclusion that utilizing software to aid rescuers in detecting victims through technology saves time and reduces the manpower required to assess aerial images.11,17,19,21,24
The studies also showed that the automation software and approaches themselves differ considerably. Weldon and Hupy 19 describe software that uses color-based spectral signatures to identify objects. Levin et al 11 explored the use of electro-optical sensors with thermal imaging to detect humans. Airborne optical sectioning software was used by Schedl in 2 studies, in which the program combines multiple aerial images and removes occlusions or obstructions to improve human detection.17,21 Yeom 24 described a dual object detection method whereby K-means clustering software and infrared thermal imaging were used. K-means clustering is a method for image segmentation by subtracting the interest area from the background. 24 This makes detection of hot spots on thermal images easier as it includes assessment of the object size, color, and heat signature.
Piloting and Path Planning
Search planning is important, as poorly planned searches can waste time and may result in victims either being left undetected or creating a need to revisit an already searched area. 19 The literature described UAVs with autonomic and manual piloted modes. Eleven studies were found that reported on the use of manually piloted UAVs,2,3,5,25,9,13-15,20,22,23 with 8 studies describing autonomous control systems that also had the capability of switching over to manual pilot mode if required.8,10,12,17-19,21,24
Autonomous modes may free up hands as the UAV can be preprogrammed to navigate and fly a specific route or pathway autonomously while continuously capturing aerial images of the area it has covered.8,21 Five studies featured this type of path planning as an aid to rationalize and simplify the use of UAVs by WSAR teams.8,12,17,19,21 The literature describes different flight patterns for searching an area, with the “lawnmower” style grid being the most popular as it allows good coverage of an area in the shortest time possible. The lawnmower search pattern is one in which the UAV gathers imagery by following a preprogrammed flight path in a manner ensuring that the flight paths slightly overlap. 19
Adaptive path planning differs from classical path planning. Adaptive path planning comprises an autonomous UAV with a real-time on-board classification system that can autonomously detect potential victims during the aerial search. When the classifier detects a potential victim, the UAV will automatically and adaptively plan to re-search that area, thus improving detection rate. 21
Limitations to the Use of UAVs in Wsar
Locating victims with a UAV conventionally requires the victim to be visible from the sky. Various wilderness environments exist that limit human detection from the air. These include environments that have steep or uneven terrain, vegetation, mist, and snow, which may significantly obscure the view of a victim.2,3,21 Unmanned aerial vehicles equipped with thermal imaging also have limitations, for they rely on detecting temperature differences between the environment and that of the victim. Challenges can be experienced when the victim’s heat signature becomes hidden or blurred by heat signatures emanating from the surrounding environment.17,21 Weather can also drastically change during a wilderness rescue operation. This can result in delays or an absolute inability to deploy a UAV.2,8 Images taken by the UAV in such cases may also not be optimal, especially if one has to operate at high altitudes and within foggy conditions. 3
Limited flight time is another widely acknowledged limitation to the use of UAVs in many different contexts. The literature reviewed describes different types of UAVs, each with a different flight endurance. The shortest reported flight time featured in the literature was 10 min, and the longest flight was cited as being 150 min.8,20 Flight time not only affects the time available to search but consequently also limits the size of the area that can be searched, including the number of flights required to complete a search of a predetermined area. 8 Other well-recognized limitations to the use of UAVs, not specifically highlighted in the sources, include the cost of the technologies, the availability of experienced pilots, and legislative restrictions to their deployment in certain locations that may have protected airspaces.
Conclusions
Through the scoping review, 21 sources of literature were found that focused on use of UAVs in WSAR operations. Unmanned aerial vehicles can serve as a valuable tool for WSAR teams to complete environmental risk assessments, locate victims via aerial surveillance, and restore communication systems. Their limitations are noted to include objects obscuring victims, uneven terrain, cost, battery-limited flight time, and their susceptibility to weather changes and environment related damage.
Limitations and Suggestions for Future Research
We found that the terminology used across the records with regard to UAVs was at times vague, and multiple terms exist for “drones.” Utilizing a single term for a UAV with regard to WSAR operations would be better to avoid misunderstandings as to which aerial vehicle is being used and to ensure that future searches of this nature are made easier. Another significant limitation in our view is the fact that there is not a great deal of literature around the area of WSAR specifically. Consequently, the WSAR terminology may not be universally understood. The fact that only articles written in English were included is acknowledged as a potential limitation. In addition, we noted the scientific “rigor” was possibly lacking in a certain number of the studies, either with the methodology being incompletely described or insufficient when it came to sample sizes and measuring the true value of UAVs in WSAR operations. Limited studies were found that focused on authentic WSAR operations, with many being mainly descriptive in nature. We suggest that additional research should be considered that focuses on reporting and describing the use of UAVs in actual SAR missions and how the value of incorporation of UAVs can be more scientifically quantified. There is also a need to further explore which types of SAR and/or disaster incidents would best benefit from the deployment of a UAV.
Footnotes
Acknowledgements
Author Contributions: conceptualization of the study, sourcing of the raw data, and approval of the final version of the manuscript (AP, BVT); contribution to data analysis and drafting of this manuscript based on the findings (CV-L).
Financial/Material Support: None.
Disclosures: None.