Heading Prediction for Unmanned Surface Vessel Based on Reservoir Computing

Accurate heading prediction is a key premise for path planning and collision avoidance of Unmanned Surface Vessels. Therefore, this paper proposes a heading prediction method for USV based on reservoir computing considering environment disturbance and the nonlinear term of the model. Firstly, a first-order Nomoto model with nonlinear terms of the model and environmental disturbance is identified by the sea trial data of the USV. Secondly, the sea trial data and the simulation data are used to train the reservoir computing model. Finally, the prediction results of the reservoir computing model are compared with the model simulation and sea test. The practicability of the proposed algorithm is verified by the comparative results.