Risk-Aware UAS Path Planning Based on Four-Dimensional Operation Volume Construction and Dynamic Collision Avoidance

A core challenge in uncrewed aircraft system (UAS) path planning lies in the potentially high risks caused by deviations between planned paths and actual flight trajectories in complex dynamic environments. In this paper, we study path planning through two phases of pre-flight and in-flight planning. A ground risk map is constructed by integrating real-world spatial characteristics. Based on this map, a pre-flight path planning method combining BRESENHAM and A∗ algorithms (B-A*) is proposed, aiming at balancing the path distance with the risk cost, as well as achieving path optimization. Highfidelity simulations of actual flight trajectories are performed to calculate confidence intervals, thereby constructing an operation volume (OV) that ensures the actual flight path remains within a safe envelope. During the in-flight phase, a real-time collision avoidance strategy based on time windows is proposed for dynamic environments. Alternative paths are generated via a path offset model, and collision risks with dynamic obstacles are evaluated by calculating the closest point of approach (CPA). The optimal avoidance path is selected by comprehensively weighing risk and distance costs. A reverse fitting method is designed to recover the original path smoothly after collision avoidance. Experiments using simulations demonstrate its ability to significantly reduce flight risks in complex environments while enhancing path safety and execution stability.