Research On Adaptive Sliding Mode Control Theory And Its Application

Sliding mode control is not sensitive to the change of system parameters and external disturbance,and has such advantages as easy implemented.These has made it to be one of the common control methods in the field of automatic control and widely used in practical engineering.But in the actual control system,there are many uncertain factors,such as part of the system state is unknown,the external disturbance is unknown,and so on.We unable to describe the exact mathematical model to control system with uncertain factors,and there is no accurate mathematical model,so that the system can not be well controlled.Otherwise the discontinuous switch character in essence of the sliding mode variable structure control method,this leads to the limitation in the practical application of sliding mode control,and hinder the further development of sliding mode control.In this paper,we use the extended state observer to observe the performance of the unknown factors,and the universal approximation property of RBF neural network,combining them with sliding mode control to estimate the uncertainties in the system,and the effect of uncertainty on the system was eliminated,improved the quality of sliding mode control.The main contents of this paper are as follows:(1)The development background,development course and current situation of sliding mode control method are described.The sliding mode control is introduced,the basic principle and dynamic performance of sliding mode control are analyzed.The shortcomings of sliding mode control are pointed out and the root cause of the problem of sliding mode control is discussed in detail,several common methods to solve the chattering problem of sliding mode control are summarized.(2)The combination of extended state observer(ESO)and sliding mode control is studied.The method uses extended state observer to estimate the unknown states and uncertainties in the system and feedforward compensation for the system.The uncertainties of the system state are offset and eliminate the limitation of the system state.At the same time,the parameters of sliding mode controller are adjusted adaptively by using the gradient descent method.By using saturation function instead of sign function,the chattering problem in sliding mode control is weakened and the stability of the system is proved by the Lyapunovmethod.Simulation results show that the control method has good control effect on the uncertain system,and has strong anti disturbance ability.(3)RBF neural network is used to approximate the uncertain terms of the system by.First of all,when the control input of the system is limited,use RBF neural network is to approximate the control input constraint and compensation for control input and the adaptive law is designed to ensure the system is asymptotically stable.When the larger initial error is used,the RBF neural network has better compensation ability to control the input limited part.Secondly,when the system is unknown,the RBF neural network is used to approximate the unknown term and the adaptive law is designed to ensure the system is asymptotically stable.Simulation results show that the method is able to control the system better when the system is unknown.