Trustworthy AUV Inspection of Critical Underwater Assets via Risk-Aware Coverage Path-Planning

In the context of underwater robotics, the ability to perform efficient and trustworthy coverage of unknown environments is crucial for tasks such as inspection, monitoring, and exploration. This paper presents a novel risk-aware coverage planning framework designed for autonomous underwater vehicles (AUVs), which integrates a global planner based on a modified Rapidly-exploring Random Tree (RRT) algorithm with a local planner that ensures terrain following and obstacle avoidance. The global planner balances exploration efficiency and environmental risk through adjustable parameters, enabling the generation of informative and safe paths in complex bathymetric scenarios. The local planner leverages potential fields and elastic band optimization to refine the trajectory and adapt to dynamic obstacles and terrain constraints. The framework is validated using realistic underwater maps and performance metrics such as area coverage and risk exposure. Results demonstrate that the proposed approach achieves robust and adaptive coverage in challenging environments, providing a reliable tool for trustworthy autonomy in subsea operations.