Neuro Adaptive Integral Sliding Mode Control based on Composite Learning for Path-Following of an Underactuated Underwater Vehicle: Blucy

This paper presents an application of an intelligent control system for an underactuated underwater vehicle called Blucy, developed for non-invasive underwater monitoring. During the monitoring, it is crucial to maintain an attitude towards the target and stay on the survey’s path to improve the collected data quality, this highlights the need for developing a sophisticated guidance and control system. Specifically, intelligent controls demonstrate adaptability in challenging environments, such as sea current disturbances, and handle both unmodelled and nonlinear dynamics of the system, like the presence of the fiber optic cable during remotely operated inspections. In this work, a robust path-following algorithm is developed. A line-of-sight guidance is used to tackle the problem of under actuation, which outputs a virtual input to the control system. A composite error learning based methodology is implemented to design a control system. An integral sliding mode control that uses a radial basis neural network to estimate the un modelled dynamics and uncertainties is developed. In addition, a disturbance observer is designed to approximate the external noise and the error made by the neural network. Furthermore, a state estimator is implemented whose error, along with the tracking error made by the controller, is used in training the neural network and disturbance observer, regarded as so-called composite error learning, which enhances the learning process, making the controller more robust against disturbances and uncertainties. The efficiency and performance of the proposed control methodology in following the desired path are studied through simulation. © 2024, Institute of Marine Engineering Science & Technology. All rights reserved.