Artificial intelligence for unmanned aircraft systems in rescue services: Challenges and possibilities

The report is a literature review of artificial intelligence (AI) with its possibilities and challenges for unmanned aircraft systems (UAS) in Swedish rescue services. UAS can rapidly gather data from potentially hazardous locations. AI can process and analyse heterogeneous data in ways that significantly exceeds human capabilities. The report is intended for decision makers, engineers, and staff responsible for technology and method development within rescue services and crisis response. The literature review shows that AI applications for UAS in many ways contributes to capabilities of rescue services. AI can support UAS technical functions for navigation, computational efficiency, image processing and coordination of several UAS. Further, AI can support specific crises by detection of fires, flooding, landslides, humans, life signs, and sounds, as well as create temporary communication networks. The literature review also shows that application of AI creates new vulnerabilities that attackers may utilise. The vulnerabilities means that UAS navigation, automated decisions and analyses may be interfered or distorted. Attackers may also utilise AI for cyberattacks against UAS. Further investigations of operational needs and countermeasures against AI-vulnerabilities are required to utilise the results of the literature review.