Neural network sliding mode control under new reaching law and application

In order to fundamentally suppress the chattering resulted from high frequency switching of sliding surface, a new reaching law is designed on the basis of index reaching law integrated with hyperbolic tangent function, after analyzing the idealized approaching movement of sliding mode, which makes approach velocity associated with the distance to the sliding surface so as to achieve adaptive adjustment, so that the switching become gently, and chattering will be weakened thoroughly; Moreover, neural networks has been used for estimating and compensating effectively against the system model uncertainty and disturbance outside; Finally, the two methods has been combined to design sliding mode control law with Backstepping. Applying the method in designing ship course controller, and carrying out simulation whose results show that the control law can neutralize the interference of ship and outside, effectively track expected heading, and inhibit the buffeting for protecting servo.