A Path Planning Method for Unmanned Rescue Collaboration System
Zhang Xiaolin, Ping Wang, Yu Qi Han · 2024
Traffic accidents seriously endanger human life and property safety, and an efficient rescue system can effectively reduce losses. When an accident occurs, congestion often occurs in the surrounding area, which significantly affects the perception and planning capabilities of rescue vehicles. The article proposes a new paradigm for unmanned rescue systems, allowing an Unmanned Aerial Vehicle(UAV) and an unmanned ground vehicle(UGV) to collaborate to move quickly through congested roadways. The UAV uses its viewpoint advantage to identify vehicle and road information, evaluate passable ranges, and generate maps for UGV path planning. Specifically, the UAV recognizes vehicles based on the YOLO algorithm and extracts lane lines by Hough transform to generate maps. An improved rapidly exploring random tree(RRT) algorithm is used for global path planning for the UGV. Numerous experiments on virtual and real environments have validated the effectiveness of the paradigm.