Researching Of Scheduling And Control Co-design Based On Variable Sampling Period In Networked Control System

Increasingly complex industrial control systems objectively demand networked control system which is the inevitable trend of control field. In this paper, the researching work is from the aspect of network scheduling and control co-design in networked control system. To decrease the bad effectiveness of the control system's stability and the control performance which are caused by network-induced delay, etc, the main research contents are as follows.(1)Learning of the simulation platform, the function modules and system implementation processes which based on the TrueTime toolbox have been carefully studied in this paper. The TrueTime toolbox can comprehensively reflect both control properties and network transmission characteristics. DC motors were choosen as the controlled objects, a multi-loop networked control systems were built up based on the platform which was a foundation for experimental study in this paper. And two kinds of classic scheduling algorithms-RM and EDF were analyzed.(2)For resource-constrained networked control system, a variable sampling period scheduler based on Levinson-Durbin algorithm was proposed. The scheduler real-time and online monitored the network resources, it achieved dynamically allocation of the resources by adjusted the sampling periods according to the network resources demanding of the control loops. Experimental results showed that the scheduler could ensure the stability of the control system. And most of the time, network-induced delay of the system owned the variable sampling period scheduler was less than the system with the fixed sampling periods in the same environment, this illustrated that the scheduler could effectively reduce network-induced delay.(3)To ensure the stability of the NCS that works in the uncertain and variable loads environment which means besides the control system nodes, there are other applications using the network resources uncertainly, this paper also presented a strategy of variable sampling periods based on support vector machine that faced the multiple-input and multiple output systems. Comparing with the former research, the strategy has considered the couping of the sampling periods and it is close to the real system. Experimental results showed that comparing with the fixed sampling period system, the strategy made the networked control systems which worked on the condition of dynamic loads always operate stable. The system runned 10s, the total IAE of the fixed sampling periods' system was infinite, while the system with the variable sampling period strategy, total IAE was 9.791, this demonstrated that the strategy can improve the overall control performance, besides that, it also can raise the schedulability of the network.