Efficient Perception and Obstacle Avoidance Flight of UAVs in Dynamic Dense Environments
Youshen Lin, Zhijun Meng, Jiachi Ji, Zichen Wang, Weiqi Gai · 2024
Safe flight in dynamic environments is crucial for the application of Unmanned Aerial Vehicles (UAVs), Existing methods lack precise classification criteria and overlook the limited field of view. Therefore, this paper introduces an efficient dynamic obstacle perception and avoidance method suitable for UAVs with limited vision in dense dynamic environments. It includes effectively distinguishing the velocities of obstacles to enhance point cloud segmentation and motion estimation. Moreover, we have designed efficient trajectory planning and active perception strategies considering the limited field of view. We have validated the superiority of the algorithm in both simulated and real-world experiments, confirming their effective environmental perception and obstacle avoidance capabilities in dynamic environments.