Risk-Aware Obstacle Avoidance Algorithm for Real-Time Applications

Researchers have developed a new risk-aware navigation algorithm that significantly improves safety for autonomous surface vessels. The hybrid system creates probabilistic risk maps of obstacles and uses a modified RRT* planner with three optimization modes to generate smooth, collision-free paths. In tests with static and dynamic obstacles, it outperformed conventional LIDAR and vision-only approaches, demonstrating superior operational safety and the ability to dynamically adapt to changing environmental risks in real-time.