Adaptive neural networks trajectory tracking control for autonomous underwater helicopters with prescribed performance. (15th November 2022)

• Record Type: Journal Article
• Authors: Wu, Zheyuan; Wang, Qing; Huang, Haocai
• Is Part Of: Ocean engineering. Volume 264(2022)
• Publication Date: 2022-11-15
• Subjects: Autonomous underwater vehicle -- Trajectory tracking -- Prescribed performance -- Preset time performance function -- RBF neural Network; Ocean engineering -- Periodicals; Ocean engineering; Periodicals; 620.4162
• DOI: 10.1016/j.oceaneng.2022.112519
• Languages: English
• ISSNs: 0029-8018

Abstract

Abstract: An autonomous underwater helicopter (AUH) is a type of autonomous underwater vehicle (AUV) that is capable of fixed-point hovering, precise and free take-off, and landing. It is effective for underwater ultra-mobile tasks, including mobile observation networks, resource exploration, and data connection. This paper aims to investigate an AUH trajectory tracking controller based on the prescribed performance method, considering the influence factors such as current disturbance, modeling uncertainty and thruster faults. A novel preset time performance function is designed in which the stable terminal time of the system can be specified explicitly, and the convergence rate of the system's dynamic process may be altered by adjusting the parameters, making it more intuitive for the designer. A speed observer was introduced to address this problem of measuring the required state information from the AUH's sensors in practice, which was enhanced using a radial basis function neural network (RBFNN) to approximate external perturbations and uncertainties. The stability of the closed-loop AUH system is proved by using Lyapunov theory. The feasibility and effectiveness of the proposed algorithm proposed approach were eventually verified by two sets of simulations for different thruster fault forms. Highlights: A observer is introduced to observe velocity information, and the relationship between the transformation error and the observed velocity is established by the backstepping method. The proposed new performance function can not only explicitly set the stable terminal time, but also limit excessive overshoot and control input in the initial period. The thruster faults are incorporated into the general uncertainty, and RBFNN is introduced to estimate it, which enhances the effectiveness compared to the conventional LSO.