A Decentralized Multiple MAV Collision Avoidance Trajectory Planning Method

In this paper, a decentralized algorithm for solving combinatorial optimization problems is used to solve the collision avoidance trajectory planning problem of multiple micro air vehicles (MAVs). According to the algorithm, the MAV avoids generating mutual interference trajectories through negotiation with its neighbors. The MAV and its neighbors transfer trajectory schemes to each other, extract track points from them, and construct virtual obstacles at these points. RRT* algorithm is used to generate trajectories to avoid these virtual obstacles to achieve the purpose of collision avoidance. In order to improve the efficiency of conflict resolution, we proposed a method based on the maximum clique algorithm to solidify the trajectory schemes of some MAVs in the system and only modify the trajectory schemes of other MAVs. We tested the effectiveness of the algorithm in a simulation environment and successfully avoided collisions in all tests.