PWM-driven model predictive speed control for an unmanned surface vehicle with unknown propeller dynamics based on parameter identification and neural prediction

• Published in: Neurocomputing (Amsterdam) Vol. 432; pp. 1 - 9
• Main Authors: Peng, Zhouhua, Meng, Chengcheng, Liu, Lu, Wang, Dan, Li, Tieshan
• Format: Journal Article
• Language: English
• Published: Elsevier B.V 07.04.2021
• Subjects: Electric motor; Model predictive control; Neural predictors; Surge speed tracking; Unmanned surface vehicles; Model predictive control; Electric motor; Neural predictors; Unmanned surface vehicles; Surge speed tracking
• ISSN: 0925-2312; 1872-8286
• DOI: 10.1016/j.neucom.2020.12.036

This paper addresses the surge speed tracking of an unmanned surface vehicle (USV) subject to unknown surge and propeller dynamics. A two-phase on-line identification and control strategy is proposed for designing a speed tracking controller without any a priori knowledge of the model parameters in surge dynamics, propeller and drive motor. In the identification phase, an adaptive parameter estimation law is used for identifying the unknown parameters in the surge speed control system. Two-layer filters are employed to assure the convergence of estimation errors in the first learning phase. In the control phase, a pulse-width-modulation-driven (PWM-driven) adaptive model predictive speed control law is proposed where neural predictors are used to estimate the identification errors and unknown sea loads based on input–output data. The stability analysis of two neural predictors is proved on the basis of input-to-state stability. Simulation results are provided to demonstrate the efficacy of the proposed end-to-end surge speed tracking of the USV without any a priori knowledge of the surge and propeller dynamics.