Robust trajectory tracking of surface vessels in ice floe sea state via discrete integral sliding-mode control and Gaussian process regression

By considering the disturbance caused by ice floes in polar regions, the trajectory tracking control problem for uncertain unmanned surface vessels (USVs) is investigated in this paper. USVs for trajectory tracking missions in polar regions are required to not only overcome common disturbances and perturbations such as model uncertainties and environmental disturbances caused by winds, waves and currents, but it must also consider the stochastic resistance generated by ice floes. However, studies on the stochastic model of ice floes resistance on USVs are insufficient, making it difficult to a design tracking controller. This paper proposes a discrete integral sliding-mode control (DISMC) with a disturbance observer based on Gaussian process regression (GPR) technique, which could steer uncertain USVs to track predefined trajectories under disturbance without knowing its upper bound. Compared to the existing methods for USV control, (1) to the best of our knowledge, this study is among the first to address the trajectory tracking control problem of USVs in ice-floe sea conditions; (2) a novel fully data-driven disturbance observer is proposed that approximates the mean and autocorrelation function of the lumped uncertainties without requiring prior knowledge about the stochastic ice resistance; and (3) a novel DISMC given the autocorrelation function of uncertainties instead of the uncertain upper bound is proposed and proved to be stable with a probability of 1. The proposed method offers a significant approach for controlling USVs in ice-covered sea areas.