Reciprocal guidance approach for suspicious collision point identification in a multi-agent unmanned aerial vehicle scenario

摘要

In a multi-agent unmanned aerial vehicle scenario, identification of the suspicious collision point is a crucial part of the collusion avoidance algorithm. This paper presents a novel suspicious collision identification approach, which is based on the assumption that the scenario participants intentionally guide themselves towards mutual interception. This reciprocal guidance approach enables the development of a closed-form solution for the expected miss distance between the agents, thus enabling the implementation of the approach in real-time systems. The algorithm incorporates agents' maneuvering capability restrictions, thus basing its decision on feasible maneuvers. Simulations demonstrated the ability of the reciprocal guidance approach to identify collisions in advance and exemplified its robustness to circularpath engagements and variable-speed agents. (c) 2021 Elsevier Masson SAS. All rights reserved.