Trajectory Planning via Region-Constrained Searching and Path-Based Collision Avoidance

Trajectory planning is one of essential items in the application of UAVs aiming to generate feasible trajectories. In this paper, we propose a refined trajectory planning scheme, which utilizes information of obstacles to achieve faster initial path generation, and uses information of front-end path for collision avoidance in back-end trajectory optimization. We test the refined path-searching method in several scenarios to certify its effectiveness, and design a simulation environment to test our trajectory planning scheme. The result shows that our scheme reaches good performance in collision avoidance and dynamic feasibility with far less time.