Research on a Stochastic Collision Avoidance Method for Unmanned Surface Vehicles Based on the Integration of IA* and IDWA Algorithms

To address the issue of collision occurrences or path planning failures when Unmanned Surface Vehicles (USVs) employ the $A^{*}$ algorithm in environments with random obstacles, this study introduces a stochastic collision avoidance method that integrates an improved A* (IA*) algorithm with an improved dynamic window approach (IDWA). Within the framework of the IA* algorithm, the selection strategy for search points and the evaluation function are optimized to enhance the efficiency of the A* algorithm’s search capabilities. Subsequently, a redundancy point deletion strategy is proposed to eliminate redundant nodes within the path and implement local planning between every two adjacent nodes using a dynamic window approach (DWA) with a modified evaluation function. This ensures real-time stochastic avoidance of obstacles on the basis of a globally optimal path, enabling the USV to successfully reach its destination. Experimental results demonstrate that the integration of the IA* algorithm with the IDWA surpasses traditional hybrid algorithms in terms of algorithm efficiency, path length, and scenario comparison.