Research On CPS Network QoS Control Strategy Based On Feedback Scheduling

As a kind of new intelligent complex system, Cyber Physical System(CPS) is the combination of the technologies of computing, communication and control. CPS is a deep integration of the information resources and the physical resources, it will follow the computer, the Internet, and once again set off a wave of the information industry. It will profoundly change the way which humans interact with the information world and the physical world. Because of its depth of integration, high complexity and large amount of information, etc., it must meet the requirement of OoS in network, in order to ensure the stability of the system. In the current research, the research for the QoS of CPS network is less.Firstly, we analyzed the related concepts and the research status about CPS, compared the differences between CPS and related technologies, discussed temporal characteristics of CPS and several key factors which influence the CPS network performance. After summarizing the temporal characteristics of each control loop, we further analysised the influence of sampling period on control performance of system and network performance. On the basis of above analysis, we proposed a scheduling strategy based on feedback scheduling, designed the network feedback scheduler, and real-time revise the sensor's sampling frequency online, for ensuring the stability of the network. The feedback scheduler includes the network monitor, the controller and the actuator for QoS. The monitor samples the network bandwidth and the rate for error transmission in equal intervals. The controller uses the sample of bandwidth to estimate the amount of network bandwidth's adjustment. The actuator calculates the available network bandwidth in the next cycle based on the adjustment amount and the rate for error transmission, thus allocates the bandwidth to each control loops which have different requirments, then calculates the sensor's sampling frequency based on the allocated bandwidth, eventually maintains the system's stability.Secondly, we propose a detailed implementation for the controller of QoS in the feedback scheduler with the utilization of the PID control and the fuzzy PID control. We introduce the PID control and the fuzzy PID control in the first, then we creat two control models in Matlab /Simulink toolbox, and complete the prediction of the bandwidth of the next cycle at last.Finally, this paper designs a simulation example for CPS. In this example, the system has three network control loops, they share the limited network bandwidth. By comparing the proposed feedback scheduler with the traditional scheduling algorithms-RM and EDF,it shows the proposed feedback scheduler's feasibility and superiority.