Multi-Parameter Scheduling Of Networked Control Systems Based On Fuzzy Neural Networks

Network Control Systems(NCS)is a real-time feedback control system that uses a network as a transmission medium to connect sensors,controllers and actuators in different locations and control nodes to form a closed loop.The network control system has received extensive attention and application due to its high reliability,easy expansion,strong anti-interference ability,and relatively low operating and maintenance costs.However,due to the introduction of the network as the transmission medium,it has caused many problems that need to be solved.In order to find a solution to these problems,it need to based on the idea of control and scheduling collaborative design,and studies the optimization of system control algorithms and the rational scheduling of network resources.In this paper,by studying the data transmission of multiple control loops in networked control systems,consider the QoS(Quality of Service)and QoP(Quality of Performance)of the network control system,a fuzzy neural network scheduling algorithm based on multiple parameters is proposed.Firstly,the error of the control loop,the error rate of change and the network transmission delay are taken as input parameters.The fuzzy rule given by the expert experience is the training sample,and the fuzzy neural network is trained and the fuzzy control rule is memorized.Through the trained fuzzy neural network,the network demand of each control loop can be dynamically obtained.Then,using the idea of real-time task scheduling,according to the idle time of each control task,the relative deadline and the criticality,the urgency parameter of the task is obtained through a linear weighting algorithm.Finally,combined with the urgency parameter and network demand parameter of the control task,the dynamic weight algorithm is used to comprehensively derive the priority of each loop and adjust it online before the next sampling time arrives.Matlab's TrueTime tool is used to build and simulate the network control system model,and the nntool tool is used to train the fuzzy neural network.Compared with the EDF scheduling algorithm,system performance has been greatly improved,indicating the effectiveness of the algorithm.At the same time,in this paper,under the limited network load,a scheduling algorithm for sampling and control collaborative design is proposed to ensure the stability and schedulability of the control system.The current network utilization is calculated by calculating the real-time control information of each loop in the current network,and the available bandwidth of the next cycle is predicted.Then,using the dynamic weighting algorithm above,the bandwidth resources of each loop are dynamically allocated according to the obtained weight parameters,and the sampling period of each loop is adjusted online,so that the control performance between the loops is relatively balanced.At the same time,it is considered that the overall performance of the network control system is not only affected by the network scheduling and sampling period,but also receives the influence of the control algorithm of each control loop controller.In this paper,a fuzzy PID controller is designed to dynamically adjust the parameters of the PID controller according to the input error and error rate to adapt to the changes of network uncertainties.The system can obtain better control performance on the basis of making full use of network resources.Through the simulation analysis of Matlab,the combination of cooperative scheduling algorithm,multi-parameter fuzzy neural network scheduling algorithm and EDF algorithm has improved the control performance of the system,which proves the effectiveness of the cooperative scheduling algorithm.In summary,this paper starts from the scheduling algorithm of the network control system,and designs the algorithm and simulation to ensure the stability and schedulability of the control system under different conditions,which has certain practical value.