COLREG-Compliant Obstacle Avoidance and Motion Planning for USVs Using Behavioral Grid Map and DRL

Unmanned Surface Vehicles (USVs) are critical in maritime applications, requiring robust motion planning in dynamic and uncertain environments. Traditional algorithms like Velocity Obstacles (VO) and Dynamic Window Approach (DWA) face limitations in adaptability and efficiency, especially in highly dynamic scenarios. This paper presents a novel framework combining behavior grid maps and Proximal Policy Optimization (PPO) to address these challenges. Key contributions are the framework combining behavior grid maps and PPO for USV motion planning in dynamic environments, ensuring adherence to COLREGs, and the validation of the proposed approach through extensive simulations.