Composite Integral Sliding Mode Control with Neural Network-based Friction Compensation for A Piezoelectric Ultrasonic Motor

In this paper, a neural network-based (NN-based) integral sliding mode control approach is presented for a piezoelectric ultrasonic motor. The precision motion performance of the motor can be adversely affected in the presence of nonlinearities including friction and disturbance, as well as parameters uncertainties. Integral sliding mode control is effective in dealing with uncertainties and disturbances. To achieve better performance on tracking desired motion trajectories, a neural network structure with modified jump basis functions are used to model and compensate the discontinuous friction in the motor control systems. This structure can approximate friction with high accuracy but require few NN nodes. Stability of the proposed control strategy is analyzed. The simulation studies are provided to demonstrate the precise tracking performance of the proposed control scheme.