The Research On Sampling Period In Networked Control Systems

Network Control Systems(NCSs) is a complex control system with communication network and control system into together, forming a whole distributed real-time feedback closed loop control system, where all the information in the senor, actuator and controller in different geographic locations are transmitted through the communication network. For NCSs, the sampling period determines the frequency of the information exchange in the network, the shorter sampling period, the faster speed of data transmission, and the more feedback information controller will get, so those will lead to the better Quality of Performance(Qo P). However, too quickly of the data transmission will makes the higher network throughput, the more congested networks, the more networked load and the longer time delay, those will lead to the worse Quality of Service(Qo S); On the contrary, when the sampling period is too long, that will improve the Qo S, but that doesn't guarantee the transmission of the real-time data, and result in worse of the Qo P. So, sampling period is the connection between the Qo P and Qo S. The sampling period in NCS are studied in this thesis, and the main research works are concluded as fellow:At first, some research results and the latest progresses of the sampling period in NCS in recent years are enumerated from three aspects: constant sampling period, variable-sampling period and multi-sampling. And the scheduling method based on sampling period are also summarized.Second, based on the structure of Internet, the main composition of network delay and the main factors in time delay generation are analyzed. And the delay-dependent stability of NCS with the continuous system and the discrete system are analysis respectively using the Lyapunov theory and LMI. And an example is presented to show the effectiveness of the method.Third, according to one continuous time plant with single sampling rate and multirate sampling rate models, the corresponding of the discrete linear quadratic performance index function are obtained using discrete the continuous linear quadratic performance index function, and derive the direct relationships between the weighting matrix.Finally, considering the augmented state space description of the time-delay networked control system, according to the delay-dependent stability criteria, assuming the system delay equals maximum allowable delay bound, so the sampling period can be calculated when performance optimization function is minimal and combined with the empirical formulae and the stability conditions.