Adaptive backstepping sliding mode control based on MLP neural network for trajectory tracking of USV

This study proposes an adaptive sliding mode control (SMC) strategy based on neural networks and backstepping method for trajectory tracking control of the underactuated unmanned surface vessel (USV). The controller is decomposed into two loops of kinematics and dynamics by using the back-stepping control. In the kinematics loop, the surge and sway reference velocities of USV are designed and regarded as virtual control laws to stabilize the position errors. In the dynamics loop, the SMC is used to design the control laws. To avoid chattering of SMC, the exponential approach rate is improved by using the arctangent function, which forms the sliding mode controller with the variable parameter approach rate. The neural network based on the minimum learning parameter method (MLP) is used to approximate the uncertain terms of the model to enhance the robustness of the system and reduce the computational complexity. The adaptive laws are proposed to compensate for the approximation errors of neural networks and disturbances. By constructing the Lyapunov function, it is demonstrated the proposed control scheme can guarantee the uniform final boundedness of all signals in the closed-loop system. Finally, simulation results on an underactuated USV further illustrate the effectiveness.