Hydrodynamic Model Identification of an Unmanned Surface Vessel with Experiment Data

Unmanned surface vessels (USVs) have become a type of important marine equipment worldwide. In order to further enhance the intelligence of USVs, it is necessary to better know the hydrodynamic characteristics of USVs which can be used to predict USV behaviors. Model identification techniques are commonly applied to obtain the parameter values of a USV system. In this paper, we propose to use both the Cybership II platform and the lab experimental USV platform as the data source. Then, the least square support vector machine method with the linear kernel function is proposed to identify the system parameters. Moreover, the neural network and least squares support vector machine with the nonlinear kernel function is proposed to identify the the black box system model. The identified model is used to predict the trajectories of both platforms, and the results verify the effectiveness of the proposed model identification algorithms.