Topology-guided Online Trajectory Planning for UAV Navigation in Dynamic Environments

UAV (Unmanned Aerial Vehicle) plays a crucial role in the research and application of autonomous navigation due to its advantages of small size and flexible movement capabilities. To solve the problem of UAV navigation in dynamic environments, a topology-guided online trajectory planning method is proposed. Multiple topological paths are searched by region-splitting and the pruning algorithm, which expands the solution space of trajectory optimization and effectively avoids trajectories falling into local minima. For dynamic obstacles, propose a receding horizon guide Point Assigning algorithm and a gradient-based B-spline trajectory optimization to generate obstacle avoidance trajectory. Simulation results show that the proposed method is robust and real-time in dynamic environments.

I. Introduction

Autonomous navigation enables UAVs to generate safer and more reliable flight trajectories, significantly enhancing work efficiency and reducing labor costs in applications such as surveying, photography, and Architectural modeling [1]. As the application scenarios for UAVs transition from static to dynamic environments, their flight trajectories risk collisions with dynamic obstacles such as pedestrians, vehicles, and other UAVs. Given the limited onboard computational resources, ensuring the real-time performance and safety of UAV navigation in dynamic environments remains a significant challenge.