A Time-Varying Lookahead-Distance Method Based on Proximal Policy Optimization for the Wave Glider

The wave glider (WG) is environment-driven and susceptible to environmental disturbances, the WG may exhibit significant cross-track errors when traditional line-of-sight (LOS) guidance law with time-varying lookahead-distance is employed. To address this, a proximal policy optimization-based LOS guidance law with time-varying lookahead-distance is proposed. By dynamically adjusting the lookahead distance, the method can enhance the robustness and accuracy of the WG's path-following in unknown marine environments. Simulation and sea trial results show that compared with traditional LOS guidance law with time-varying lookahead-distance, the proposed method significantly improves the path-following accuracy of the WG under various conditions, which confirms its potential and feasibility for applications in unknown marine environments.