Research On Networked Control System Based On Generalized Predictive Control

With the vigorous development of the network technology, the network controlsystem has gradually replaced the traditional point-to-point control systems, and bewidely used in various fields of industrial production, and gradually benefit to themost of people. At the same time, the network delay, data packet loss, informationscheduling problem becomes the focus and hotspot of academic research, so the studyof the application of networked control systems will bring new challenges.In thispaper,I will study the problem of time delay in the network control system.Firstly,an improved generalized predictive control algorithm based on thetraditional GPC algorithm was proposed in this paper,which inherited the minimumvariance control. In the derivation process,it avoid the Diophantine equation solution,the matrix inverse operation, so it suitable for multi step predictive control of thenetwork control system.Secondly,according to the presence of unknown parameters in the model ofcontrolled object, the paper based on the recursive least square method, furtherderived the recursive least squares method with forgetting factor, and combining itwith the GPC algorithm to identify the unknown parameters in the model. In order tohighlight the advantages of the method, the paper also compare the model parameterswith using the recursive least square identification method gets.Then according to theproblem of network delay, the paper using the recursive least square with forgettingfactor method to predict the output in different network delay,in order to analysis theeffect of the method.Finally,considering the limitation of the generalized predictive controlalgorithm in nonlinear system, the thesis proposed combine the fuzzy control theorywith the generalized predictive control algorithm and effects it on nonlinear system,firstly analyzes the possibility of the combination of predictive control and fuzzycontrol in the basic principle,secondly, the thesis puts forward the fuzzy T-S model aspredictive model of generalized predictive control,thus forming a kind of fuzzygeneralized predictive control method. To further confirm the researching work,the paper gave the correspondingsimulation analysis.The analysis of simulation showed: The GPC algorithm withminimum variance control thought, compared with the traditional GPC algorithm, itnot only put network delay into account, but also it has the accurate forecasting output,small fluctuation, and can fastly tracking the steps.And according to the simulationanalysis of the recursive least square method with forgetting factor showed: therecursive least squares method with forgetting factor not only able to correctlyidentify the parameters of the model, and compared with the least square method, ithas the advantage of high accuracy, short transition time and so on.in the simulate ofthe nonlinear networked control system,it showed that the nonlinear networkedcontrol system, which using the orthogonal least square method can well establish theforecasting model,identify model parameters,and can realize the accurate tracking ofthe set square wave signal.