Continuity Path Planning for UAV with Enhanced Douglas-Peucker Algorithm

This paper proposes a continuous path planning method, where Deep Q-network (DQN) is employed for initial grid map path planning, followed by the enhanced Douglas-Peucker (DP) algorithm to address obstacles, and culminating in the generation of discrete path points using curvature-continuous Bézier curves. First, for grid maps, an unmanned aerial vehicle (UAV) path planning method is designed, utilizing the DQN algorithm to generate a discrete set of path points. Second, an enhanced DP algorithm is used to filter the discrete path points in an environment with obstacles. Then, the filtered set of path points is smoothed using Bézier curves with continuous curvature to conform to the flight physical characteristics of the fixed-wing UAV. Finally, the effectiveness of the proposed path planning and smoothing algorithm is verified through numerical simulations and comparative simulations in a grid map environment with random obstacles.