RBF Neural Network Integral Sliding Mode Control for Three-Dimensional Path Following of Underactuated AUV

For the three-dimensional path following problem of an underactuated autonomous undersea vehicle(AUV) under the upper bound of unknown parameter perturbation, a spatial path-following controller is proposed based on the radial basis function(RBF) neural network integral sliding mode to reduce chattering caused by external disturbance and sliding mode. The position error stabilization of the path following is converted into the stabilization of the line of sight (LOS) angle error by introducing the Serret-Frenet local frame and the LOS guidance law. And the kinematic “virtual control law” is designed for stabilization of the LOS error based on the Lyapunov direct method. Then, the dynamic path-following controller is developed by introducing the RBF neural network and the integral sliding mode surface. The stability of the designed path-following controller is analyzed by using Lyapunov stability theory. Simulation results show that this path-following controller can effectively follow the three-dimensional straight path, improve the control accuracy, reduce the chattering, and display certain robustness to the external disturbance without losing its rapidity.