UAV Collision Avoidance Based on Varying Cells Strategy

This paper presents a trajectory planning strategy for collision avoidance in unmanned aerial vehicles. First, a varying cell strategy is proposed to integrate aerodynamic constraints into trajectory planning. Basic avoidance actions in the varying cell strategy are adapted accordingly to go through different cells, enabling more flexible avoidance maneuvers. Second, given the limited decision-making time involved, offline and online path planning methods are developed to increase the convergence rate. Finally, Monte Carlo simulations demonstrate that the proposed method satisfies aerodynamic constraints, while both the convergence and collision avoidance rates are better than that achieved by fixed cell-based methods.