Topological Risk-Based Path Selection in Dynamic Environments
Erpei Xu, Chengpu Yu, Yixuan Liu · 2024
Path selection for autonomous drones in dynamic scenes poses a significant challenge. The changing environmental topology, affected by moving obstacles, can cause previously established path prone to be unsafe. This article introduces a method for selecting safe topological paths in dynamic environments without the need to model moving obstacles. This method involves generating several smooth topological paths and then choosing the safest one through calculating the risk associated with each path on a particle dynamic map. Furthermore, a replanning strategy is proposed in dynamic scenarios to proactively address upcoming obstacles, enhancing navigational safety. Finally, the proposed method is validated in both simulated and real-world environments.