Tracking Performance Limitations Of Networked Control Systems With Communication Constraints

With the fast development of computer, electronic and network communication technique, networked control systems (NCSs) arise at the historic moment. NCSs consists of geographically distributed sensor, controller and actuators, which are connected by a bandwidth limited digital communication network, and it is a kind of spatial distribution of real-time feedback closed loop control systems. NCSs do not have shortcomings of traditional control systems the high cost, such as poor reliability, difficulty in diagnosing and maintaining fault, and so on. Due to the advantages including shared information resources, easy to maintain and extend, high reliability, high efficiency, flexibility, and so on, NCSs have been widely used in industrial automation, distributed mobile communications, unmanned aerial vehicle and many other fields.Being two important topics in control fields, the stability and performance of control systems have attracted much attention all the time. Stability is the premise for system working normally. While the study of performance of control systems aim to achieve performance limitation of stable systems by designing controllers. Currently, relevant results have revealed that the tracking performance limitations of control systems are determined by the essential characteristics of reference signals and the plant. However, with networks introduced into control systems, the stability and performance of NCSs have unprecedented challenges, and thus this thesis focus on the study of tracking performance limitations of NCSs with communication constraints. The research content of this dissertation is outlined as follows:This dissertation investigated the tracking performance limitations of NCSs with channel input power constraints, and under the effects of input disturbance and Gaussian white noise. The exact expression of tracking performance limitations of NCSs has been obtained, which show that the tracking performance limitations of NCSs is closely related to the essential characteristics of the plant such as locations and directions of nonminimum phase zeros and unstable poles, and reveal that the tracking performance may be badly degraded by input disturbance and Gaussian white noise in some degree. In anther aspect, the tracking performance of NCSs becoming better with the increase of the allowed channel input power.Also, the tracking performance limitation of NCSs under effects of packet dropouts is investigated. The packet dropouts are modeled as a binary stochastic process. With unity feedback and two-freedom-of-degree control structure, a lower bound and an exact expression of tracking performance have been obtained respectively. The results reveal that compared with the unity feedback, the two-freedom-of-degree control can eliminate the negative effect of noise on tracking performance of NCSs. In addition, due to the frequentlyly occurs of packet dropouts, that is, with the increase with packet dropouts’ probability the tracking performance will be badly degraded, even makes the NCSs unstable.This dissertation studied the tracking performance limitations of multi-input multi-output NCSs with encoder-decoder. Based on different locations of encoder-decoder (feedforward channel and feedback channel), two-freedom-of-degree controller is adopted to track the given stochastic reference signal. The results show that the tracking performance is tightly dependent on the characteristics of the plant and encoder-decoder, and well-designed encoder-decoder can improve the tracking performance limitations of NCSs.Moreover, the tracking performance of discrete-time NCSs with channel input power constraints, and the regulation performance of NCSs with control energy constraints were studied in this dissertation. A new performance index for the tracking performance is proposed. Based on the optimization method with the inequality constraints, a lower bound of tracking performance and exact expression of regulation performance are obtained. In addition, the design methods of controllers of the tracking performance with channel input power constraints and regulation performance of NCSs with control energy constraints have also been obtained.The stability problem of NCSs with packet dropouts, input disturbance and additive Gaussian white noise is studied. Base on the all-pass factorization of transfer function matrices, Youla parameterization of controllers and optimization method with the inequality constraints, the minimum values of signal-to-noise ratio guaranteeing the stability of NCSs are derived. The results quantitatively reveal how the signal-to-noise ratio performance is affected by the nonminimum phase zeros and unstable poles, which can also be understood in a different way that when NCSs is stable, the explicit relationship between communication basic parameters (signal-to-noise ratio, etc.) and communication induced parameters (packet dropouts, etc.), which can give some guidance for the design of communication channel.For general non-square plant, this dissertation studied the tracking performance limitations under the effects of input disturbance and Gaussian noise. The form of Youla parameterization of controllers for non-square plant model has been determined. The explicit expression of tracking performance limitations have been obtained based on inner-outer factorization.Finally, a summary has been proposed for all discussions in this dissertation. The further study and research works are presented for the network systems.