PID Control For Several Classes Of Networked Systems With Incomplete Information

This thesis is devoted to the proportional-integral-derivative(PID)control problem is investigated for several classes of networked systems with incomplete information.More specifically,the causes for incomplete information mainly include some commonly encountered network-induced phenomena(fading measurements and time delays),randomly occurring cyber-attacks,and some data transmission protocols(event-triggered protocol and Round-Robin protocol)which are introduced to mitigate the communication burden.The addressed PID control problems cover the PID output-feedback control problem as well as the observer-based PID control problem and,on this basis,the design problem of proportional-integral observer(PIO)for networked systems is dually studied.Moreover,some novel concepts and performances in mean-square sense are adopted/proposed to cope with the problems mentioned above,such as the exponentially mean-square input-to-state stability,the mean-square input-to-state stability,the mean-square security,et al.Specifically,the content of this thesis can be divided into three parts.In the first part,the focus is on the PID output-feedback control problem with the effect of incomplete information.The PID output-feedback control problem is investigated for a class of linear discrete-time systems with incomplete information.The fading measurement phenomenon is take into account for the PID output-feedback controller design problem.By means of Lyapunov stability theory and stochastic analysis technique,sufficient conditions are obtained to guarantee the exponentially mean-square stability and the prescribed H? performance and,on the basis of such conditions,the synthesis issue of the PID output-feedback controller is subsequently discussed,where the LMI-based orthogonal decomposition technique is introduced to facilitate the controller design.Moreover,based on the similar analysis method,a PID output-feedback controller is developed for a class of linear discrete-time systems such that both the exponential stability and the l2-l?performance are guaranteed in the presence of time-varying delays.In addition,under the static event-triggered protocol,the PID output-feedback controller,which meets the H? performance constraints,is designed.The second part studies the observer-based PID control problem in cases of incomplete information(which are mainly caused by randomly occurring cyber-attacks,network-induced phenomena and static/dynamic event-triggered protocol).The PID output-feedback controller is firstly designed for a class of networked control system under the static event-triggered protocol.Next,for a kind of linear discrete time-delay systems subject to cyber-attacks,the observer-based PID security control problem is considered where two sequences of Bernoulli distributed random variables with certain probabilities are adopted to characterize the random occurrence of the cyber-attacks.A theoretical framework is established for examining the exponentially input-to-state stability in mean-square sense,and the desired security level is then achieved.Subsequently,an upper bound of the quadratic cost criterion is obtained and the explicit expression of the desired PID controller is also parameterized.Furthermore,the observer-based PID security control problem is addressed for a class of linear discrete-time systems under dynamic event-triggered protocol and deception attacks.A dynamic event-triggered mechanism,whose threshold parameter is dynamically adjusted according to a certain rule,is exploited to modulate the transmission of data packets with hope to effectively alleviate unnecessary energy consumption.The input-to-state stability of the closed-loop system is analyzed to account for the effect from the randomly occurring deception attacks and the dynamic event-triggered protocol on the system performance.An observer-based PID controller is constructed to ensure the prescribed security index of the closed-loop system.In the third part,the PID control problem mentioned above is dually extended to the PIO design problem for networked systems with incomplete information.The Round-Robin protocol is employed to schedule the data transmissions from the sensors to the observer so as to prevent the data collisions.A novel PIO is developed whose gains are dependent on the data transmission order as a reflection of the effects induced by the Round-Robin protocol scheduling.By resorting to the token-dependent Lyapunov functional,the desired PIO is designed with exponentially stable error dynamics of the state estimation and guaranteed l2-l? disturbance attenuation/resistance capacity.Gain matrices of the desired PIO are parameterized in terms of the solutions to the token-dependent LMIs that are readily solvable.