Modeling and Control of Unmanned Surface Vehicles: An Integrated Approach

This paper presents a comprehensive approach to augment the control performance of unmanned surface vehicles (USVs), addressing two core issues: dynamics modeling and control of USVs. To bolster the precision of dynamics modeling, the paper introduces a parameter identification algorithm based on the nonlinear multi-innovation least-squares method (NMILS). NMILS helps mitigate the noise influence and enhances the precision of the dynamics modeling. To further reinforce control performance, finite-time sliding mode control (FTSMC) is employed. FTSMC effectively counteracts the influence of identification errors, offering enhanced robustness against uncertainties and disturbances. The proposed techniques are validated on the Cybership I model. Simulation results revealed highly accurate parameter identification, with identified values for key parameters m11, m22, and m33 closely matching the true values. Moreover, motion prediction with these identified parameters yielded minor errors, the largest spread being in eu with a maximum value of 0.047m/s. The effectiveness of the FTSMC control strategy was demonstrated through a path-following simulation. Notably, the maximum errors for xe and ye did not exceed 0.006m and 0.15m respectively, reinforcing the precision of the proposed approach.