Predictive Control Of Networked System Under The Non-ideal Networked Conditions

With the increasing complexity of the control structure and task,and the rapid development of sensor and computer communication technology,a new type of networked control systems(NCSs)emerge as times require.Because of their advantages such as easy wiring,easy operation and maintenance,low cost,high efficiency and flexibility,the NCSs have been widely used in many fields.However,for the reason that all kinds of information in NCSs share the limited network resources,the network-induced factors are inevitable under the non-ideal networked conditions,such as time delay,packet dropout and packet disorder,which may lead to poor performance,oscillation and even instability.As a result,the research methods of the traditional control systems can not be directly applied into NCSs under the non-ideal networked conditions.It is one of the important research issues on NCSs that applying the new control theories and methods to ensure the system stability and control performance.In this dissertation,focused on the NCSs under non-ideal networked conditions,we have established the model of the systems and analyzed and controlled the systems.The main contents of the dissertation can be concluded as follows:(1)For a class of discrete-time linear NCSs with time-varying delays,we have studied the stability and conservatism of the systems.By constructing a novel augmented Lyapunov functional and using an improved finite-sum inequality to deal with some sum-terms appearing in the forward difference of the Lyapunov functional,a less conservative stability criterion is obtained for the studied systems if compared with some existing methods.Moreover,as a special case,the stability of discrete-time systems with a constant time delay is also investigated.(2)For linear NCSs with data dropout,the observer-based event-triggered predictive controller has been designed considering the unmeasurable of the states.The control scheme proposed not only reduces the amount of sampling-signal and prediction-signal transmission,saving the limited resources of networks,but also compensates the negative effects of data dropout actively.Futher,by applying Lyapunov stability theory,the design method of observer and predictive controller is given by linear matrix inequalities.(3)For nonlinear NCSs described by T-S fuzzy systems with data dropout,we have designed the fuzzy event-triggered predictive controller.Since the imperfect premise matching is well considered,the design flexibility and low cost of the fuzzy controller can be expected.A networked T-S fuzzy model is well constructed which has integrated the event-triggered communication scheme(ETCS)and the predictive control together,in which,a)an ETCS is introduced to alleviate the communication burden by reducing the number of transmitted packets;b)an fuzzy predictive controller is designed to predict future states and control signals between two successfully transmitted instants.Therefore,the data dropout induced by the networks can be actively compensated.