Integral Sliding Mode Trajectory Tracking Control for UG Based on RBF Neural Network Observer

In order to study the longitudinal trajectory tracking problem of underwater glider (UG), an integral sliding mode control based on radial basis function (RBF) neural network observer is proposed in this paper. Firstly, a hydrodynamic model of longitudinal motion of UG is constructed. Then, an RBF neural network (NN) is used to approximate the nonlinear function of UG hydrodynamic model, and an adaptive observer is designed to observe the state of the system (i.e., pitch angle and pitch velocity). In the process of parameter configuration of the neural network, the adaptive learning rate is introduced to improve the accuracy of the hydrodynamic model of trajectory tracking. Furthermore, in the design of integral sliding mode controller, improved hyperbolic tangent function is used to eliminate chattering phenomenon. The simulation results show that the control scheme designed can effectively track the control instructions of the glider with global robustness.