Anti-disturbance depth control based on the robust L1 filter for under-actuated AUV with switched linear parameter varying dynamics

• Published in: Journal of marine science and technology 2023, Vol.28 (1), p.179-194
• Main Authors: Bo, Peng, Tu, Xingbin, Wei, Yan, Qu, Fengzhong
• Format: Article
• Language: English
• Subjects: Amplitude; Amplitudes; Automotive Engineering; Autonomous underwater vehicles; Basis functions; Closed loops; Connectors; Control systems; Control systems design; Controllers; Couplings; Depth control; Disturbances; Dynamic stability; Dynamics; Engineering; Engineering Design; Engineering Fluid Dynamics; Feedback control; Linear matrix inequalities; Marine environment; Mathematical analysis; Mechanical Engineering; Offshore Engineering; Original Article; Parameters; Robust control; Stability; Underwater vehicles
• ISSN: 0948-4280; 1437-8213
• DOI: 10.1007/s00773-022-00909-2

To improve the anti-disturbance stability of the autonomous underwater vehicles (AUVs) under bounded disturbances in amplitude, we develop a robust controller based on the estimated signals generated by the filter. The AUV studied has the characteristics of time-varying delay, under-actuated, and switched linear parameter varying (LPV) dynamics. Both the filter and the control strategies, we proposed are based on the robust L 1 performance criterion, which is suitable for the amplitude-bounded disturbances in the ocean environment. In the filter and controller design, the corresponding Lyapunov–Krasovskii functional is established to verify the stability of the systems. As there exist couplings between the designed functional and the system parameter matrices, slack matrices are constructed for decouplings after validating the stability of the under-actuated AUV filtering error system and closed-loop control system. The filter and the controller are obtained in the form of parameter linear matrix inequalities (PLMIs), whose solution is infinite-dimensional matrix inequalities. The approximate basis function and gridding technique are applied to transform the filter and the controller into that of finite dimensional LMIs. The simulation has verified the effectiveness of the robust L 1 filter and anti-disturbance depth robust L 1 controller.