Robust Path Tracking and Obstacle Avoidance of Autonomous Ship using Stochastic Model Predictive Control

This paper presents a stochastic model predictive control (SMPC) approach for robust path tracking and obstacle avoidance of an autonomous ship. In real environment, state estimation error, environmental disturbance and uncertainties in the force generated by actuators contribute to the uncertainty of the model and controller. This uncertainty may make accurate modeling difficult and decrease controller performance, which may lead to a ship collision accident. SMPC is a promising solution to address the challenges associated with ship movement modeling and uncertainty. The linearized ship heading control model is used in the study, and state uncertainty propagation is performed to account for the ship's uncertainty. Chance constraints are applied to manage the probabilistic constraints for the uncertain state, and an auxiliary feedback controller is designed to overcome the uncertainty. The trade-off between conservativeness and risk is managed by a risk factor in the control design process. The proposed algorithm's effectiveness is verified through simulation results.