Abstract:
This article presents implementation aspects and experimental results for a cooperative multi-UAV team sensor and perception system. The presented on-board system is integrated in a R&D full-mission helicopter simulator used for manned-unmanned teaming (MUM-T) research in complex military search & rescue scenarios requiring field landings in uncontrolled and unsafe areas. Thus, to reduce the risk, the presented multi-UAV system is capable of performing highly-automated landing zone reconnaissance and landing point evaluation. Thereby, the presented sensor and perception management system (SPMS) incorporates knowledge on information demands regarding safe landing points and probabilistic reliability estimations of applied perceptive capabilities to derive its own course of actions for the landing zone reconnaissance. Expert knowledge is used to weight the single criteria in the multi-dimensional landing point assessment process while a perceptive trustworthiness estimation is used to handle the uncertainty of the measuring processes when fusing the reconnaissance results to assess the single landing points and to create a landing point recommendation for the helicopter crew. The presented system was evaluated in an extensive human-in-the-loop campaign with German Army Aviators. This paper presents questionnaire-based results gathered during the experimental campaign, focusing on human factors, user acceptance, and system design aspects.