UAV Path Planning in Urban Environments with Dynamic Risk-Map Generation by Vehicle and Pedestrian Perception

In this paper, we propose the UAV path planning method in urban environments with dynamic risk-map generation by vehicle and pedestrian perception. In urban environments, it is the challenging task for Unmanned Aerial Vehicles (UAVs) to generate the safety flight path such that they will not collide with the ground vehicles and pedestrians in case of accidental fall. To mitigate the collision risk, the proposed method dynamically generates the risk-map, where collision risks within specific areas are mapped based on mobility of ground vehicles and pedestrians. To perceive the mobility of them, the proposed method utilizes Vehicle-to-Everything (V2X) messages that include position, velocity, and acceleration of the ground vehicles and pedestrians. Based on the proposed risk-map, UAVs seek their flight paths that minimize the collision risk by Rapidly-exploring Random Tree Star (RRT*) algorithm. Simulation result shows that the flight path of the proposed method decreases the collision risk compared with the shortest path and previous safety path in the situation that there are 200 vehicles and 400 pedestrians in 1250 m×1050m city.