Tracking Control Of Distributed AUV System Based On Non-singular Fast Terminal Sliding Mode

With the depletion of natural resources,how to explore marine resources efficiently has become the focus of contention among many countries in recent years.Marine science and engineering is a subject which includes various complex ocean tasks,such as seabed resource exploration,sea surface target search,ocean climate observation and so on.Considering the harsh deep-sea environment and the limitation of human diving depth,autonomous underwater vehicles(AUVs)play a unique role in the deep-sea area.Distributed AUV systems consisting of multiple AUVs have a wider range of operations and higher energy efficiency,which are becoming more and more universal and applicable.In addition,considering the relative independence of each AUV,distributed AUV controller is particularly important for multi-AUV systems.However,how to ensure that each AUV in the system completes the assigned tasks reasonably and orderly,and achieves consistency tracking control has become a major problem in the field of AUV control.Based on the second-order AUV dynamic model,this paper combines the leader-follower structure with the directed graph theory to describe the communication topology of distributed AUV system and illustrate the system structure intuitively and vividly.Then,the global tracking errors which are made of absolute tracking errors and relative tracking errors are defined by combining graph theory with dynamic model expressed by Euler equation.In order to illustrate the effectiveness of the proposed algorithm,the communication of time-delay is added to the error definition,and the feasibility of the algorithm expansion is analyzed theoretically.In the design process of distributed AUV controller,different technologies are introduced to ensure the stability and robustness of the system and overcome the adverse effects of harsh marine environment.Firstly,in order to overcome the unknown uncertainties caused by hydrodynamic coefficients,hydrodynamics and other factors,as well as the internal and external nonlinear uncertainties of the multi-AUV system,adaptive control and neural network technology are used to approximate the unknown nonlinear functions.The simulation results show that the approximation property represented by the optimal weight vector can solve the problems of nonlinearity,strong coupling and time-varying parameters in the system.Secondly,the finite-time control is combined with the non-linear terminal sliding mode control,which ensures the accessibility and stability of the sliding surface and improves the fast response characteristics of the system.An improved nonsingular fast terminal sliding mode is constructed,so that the global tracking error converges to the small neighborhood near the origin in finite time along the sliding surface.Finally,the finite-time Lyapunov stability theory is used to analyze the stability of distributed AUV system,which proves that the control strategy can ensure that the following AUV has good tracking performance and anti-interference ability.Simulink module of MATLAB software is selected to build the simulation experiment,and the experiment compared with traditional controller is added.The experimental results show that the neural network controller based on non-singular fast terminal sliding mode can significantly improve the fast response characteristics and robustness of distributed AUV system,and the tracking effect of each AUV is better than traditional controller.The system has high consistency and stability,which further proves the accuracy and effectiveness of the proposed algorithm.