Autotune control algorithm based on relay feedback and adaptive neural network for attitude tracking of nonlinear AUG system. (15th April 2022)

• Record Type: Journal Article
• Authors: Hao, Jun; Zhang, Guoshan; Liu, Wanquan; Zou, Haoming; Wang, Yanhui
• Is Part Of: Ocean engineering. Volume 250(2022)
• Publication Date: 2022-04-15
• Subjects: Data driven control -- Adaptive neural networks -- Relay feedback -- Fast adaptive learning factor -- Nonlinear AUG system -- Lyapunov stability theory; Ocean engineering -- Periodicals; Ocean engineering; Periodicals; 620.4162
• DOI: 10.1016/j.oceaneng.2022.111051
• Languages: English
• ISSNs: 0029-8018

Abstract

Abstract: Due to the complexity and uncertainty of the nonlinear autonomous underwater glider (AUG) system, the control algorithms for attitude tracking of the AUG system are very difficult to directly design. In this paper, a novel autotuning control algorithm (ATCA) based on relay feedback and adaptive neural network is proposed to effectively implement the attitude tracking of the AUG system. The proposed algorithm only utilizes the online input/output (I/O) data to achieve the AUG system attitude control, ignoring the mathematical system model. The ATCA control parameters are initialized by relay feedback and adjusted online based on gradient descent algorithm with the partial derivative of the AUG system provided by adaptive neural network. Besides, in the ATCA, the fast adaptive learning factor is employed to make the AUG system respond quickly to the evolving reference trajectory. Furthermore, the complete stability of the closed-loop AUG system with the ATCA has been proven via the Lyapunov stability theory. The simulation studies illustrate the correctness the proposed algorithm. Compared with three popular data driven control algorithms, the proposed algorithm has superiority in terms of system response time, integral squared error (ISE) and integral absolute error (IAE). © 2014 xxxxxxxx. Hosting by Elsevier B.V. All rights reserved. Highlights: A data-driven algorithm named ATCA is proposed to achieve the AUG attitude control without relying on AUG system model. Relay feedback solves the problem of controller parameters initialization, and only one learning factor is manually set. The complete stability of closed-loop AUG system with proposed ATCA has been proven via Lyapunov stability theory. In comparation with MFAC, LM-PIDNN-RF and RFPID, the simulation results of the AUG indicate the superiority of the ATCA.