Neural sliding-mode tracking control for a class of nonlinear-actuators

In this paper we propose a neural sliding-mode control scheme for systems include a class of nonlinear operators. The plant is represented as a linear system preceded by a nonlinear operator, which is modeled with a Prandtl-Ishlinskii (PI) operator. In order to eliminate the nonlinearity, an approximate operator is used as a feedforward compensator. A sliding mode controller is then used to mitigate the effect of inversion error. In the existing work using neural network with sliding-mode control, the modeling of nonlinearity and compensating for it is not considered as the neural network can replace it. In contrast, we showed that adding an inverse-operator will improve the results particularly with sinusoidal references. The stability of the closed-loop system is established by Lyapunov analysis. Simulation results are presented to demonstrate the effectiveness of the proposed method.