ESHLOS guidance for neuro-adaptive path following control of underactuated surface vessels

• Published in: Ocean engineering 2022-12, Vol.266, p.112894, Article 112894
• Main Authors: Xia, Guoqing, Wang, Xinwei, Zhao, Bo, Han, Zhiwei, Ren, Zheda
• Format: Article
• Language: English
• Subjects: ESHLOS guidance law; Input saturation; Neural networks; Path following; Underactuated surface vessels
• ISSN: 0029-8018; 1873-5258
• DOI: 10.1016/j.oceaneng.2022.112894

This paper proposes a neuro-adaptive path following control scheme based on ESHLOS guidance law for underactuated surface vessels with unmeasured linear velocities. Based on two extended state observers, an ESHLOS guidance law is designed, where the guidance law not only calculates the desired heading angle and surge velocity signals, but also compensates the unknown sideslip angle. Then, neuro-adaptive path following controllers are proposed to track the desired signals via backstepping technique. The hyperbolic tangent function is applied to model input saturation nonlinearity, and adaptive radial basis function neural networks are used to offset the lumped disturbances including external environmental disturbances, model uncertainties. Besides, a high-order tracking differentiator is introduced to generate derivatives of desired heading angle, so that the computational burden inherent in backstepping technique is reduced. Based on the Lyapunov functions, it is proven that all error signals of the closed-loop system are bounded. Finally, simulation results are utilized to reveal the efficacy of the presented approach. •An ESHLOS guidance law is designed to generate desired surge speed and heading angle, and compensate the unknown arbitrary sideslip angle.•The input saturation nonlinearity is modeled by a hyperbolic tangent function, so that the backstepping technique can be applied directly in the control design.•The adaptive neural path following control laws are designed with the lumped disturbances approximated by RBFNNs.•A high-order tracking differentiator is introduced to smooth the referenced heading angle and reduce the computational complexity inherent in the conventional backstepping technique.