This paper presents an adaptive receding horizon controller for a high-speed, jet-powered catamaran Autonomous Surface Vehicle (ASV) operating in narrow and cluttered natural waterways. Using an integrated vision and laser-based sensor system for real-time shoreline classification and obstacle detection, the control reference trajectories and planning horizons are adapted based on the estimated obstacle/shoreline position uncertainty and the predicted waterway trajectory. Simulation and experimental results demonstrate the performance of the proposed shoreline detector, planner and controller on an ASV in a cluttered and narrow waterway environment.
展开▼