Sliding Mode Control For Networked Control Systems Subject To Packet Disordering Via Prediction Method

By using the technology of network,communication and computer,people connect sensor,controller and actuator in the form of network and model closed-loop NCSs.The introduction and popularity of control networks can improve the efficiency,flexibility and reliability of traditional industrial process,power system and transport system and reduce costs of the installation,the maintenance and the reconfiguration.Now this technology is widely adopted in the processes of industry,business and manufacture.In fact,there are some problems such as network induced delay,packet loss and packet disordering in transmission channel,which will degrade the performance of the system.Thus it is meaningful to find and solve the issues involved in the NCSs.Given that current research mainly focuses on time delay and packet loss,this thesis will concentrate on the problem of packet disordering.Under the circumstance of occasionally unpredictable packet disordering,this thesis predicts the packet disordering firstly,and then designs sliding mode control(SMC)for nonlinear and varying-parameter NCSs with external disturbance.The works are listed as follows:1.Considering unknown packet disordering and disturbance in NCSs,a scheme of predicting packet disordering and designing SMC with Hoo technique is proposed.Firstly,the nonlinear system is transformed into a linear varying-parameter system through the differential mean value theorem.Secondly,autoregressive moving average(ARMA)model is built to predict packet disordering directly,and then time delays are converted into the displacement of packet disordering via a new definition of packet disordering measurement.In this case,predicted packet disordering will be obtained.Next,the strategy of adopting the newest control input with a stochastic parameter leads to a linear Markovian jumping system(MJS).After this,Lyapunov function and SMC with H? controller are designed to stabilize the system.Finally,the validity of the proposed method is verified by the practical examples.2.With another different kind of nonlinear NCSs,the prediction method is modified and novel sliding surface and sliding mode controller are designed.First of all,a NCS is established with using high-precision extreme learning machine(ELM)to predict packet disordering.Then new and mode-dependent sliding surface is designed to resist the uncertainties of nonlinear parameters and vary-structure parameters.For the analysis of system stability,the thesis designs relative Lyapunov function,and finally simulation examples are given to illustrate the effectiveness of the proposed strategy.3.As for the NCS with strongly random packet disordering,the concept of packet disordering probability prediction is presented in chapter 5.Firstly,an argument system is presented with states and sampling period.Secondly,the measurement of packet disordering is presented and the statistical analysis of packet disordering is proceed.It turns out that the sampling strategy helps to reduce heavy packet disordering.Then,Markov chain prediction model is set up to predict the probability of the occurrence of packet disordering.There exist two stochastic parameters,thus the linear system will be built as a MJS with two independent Markov chains.After this,mode-dependent integral sliding surface is constructed,and sliding mode controller is designed to drive the dynamic onto the pre-specified sliding surface.Finally,simulation examples are given to illustrate the effectiveness of the proposed strategy.