Credibility-Weighted Value Learning

Trust, Geometry, and Rules: A Credibility-Aware Reinforcement Learning Framework for Safe USV Navigation under Uncertainty

Announce Type: replace Abstract: Autonomous navigation of Unmanned Surface Vehicles (USVs) that is safe and compliant with the International Regulations for Preventing Collisions at Sea (COLREGs) remains a formidable challenge in dynamic maritime environments, particularly when perception systems exhibit miscalibrated uncertainty. Existing Reinforcement Learning (RL)-based methods often falter because state-estimation errors induce unreliable belief states that mislead the value function,...