Research On The Scheduling Method Of Networked Control System Based On Fuzzy Neural Networks

Networked control system is a closed-loop feedback control system with the components such as sensors, controllers and actuators in the system connected through the network. This kind of control system is widely attention with high reliability, less connection and easy to extend, etc. But the networked control system also inevitably raises a lot of problems: network-induced delay, packet dropout and network congestion etc. These problems impact the system control performance and the stability of the system. Therefore, the study of network control system indeed has practical significance.The control performance and stability of network control system are depend on the scheduling distribute of control algorithm and network resources of the control system. Therefore, in addition to deal with related issues by designing control system, we can also consider about the aspects of scheduling algorithm. To improve the performance of the system by scheduling algorithm.This paper research on network scheduling strategy in order to improved the performance of the system. The research ideas and methods as following:According to control system of network data transmission problem, this paper proposes a multiple parameter fuzzy neural network scheduling algorithm. Firstly, training loop error and error change rate by the self-learning ability of the fuzzy neural network, and memorizing the fuzzy control rules to determine network demand of each control loop. Secondly, combining with the network demand parameters and network emergency degree, a dynamic weighting algorithm has been proposed to determine the priority of each circuit, and realize the priority of each circuit online adjustment. Finally, establish a multiloop control system simulation model with the weights of fuzzy neural network scheduling priority dynamically adjustable. The simulation results show that in the same network bandwidth utilization rate the scheduling algorithm has the better performance and stability of the system than EDF scheduling algorithm.In view of the affect of the limited network bandwidth resources for networked control system, scheduling algorithm with variable sampling period is proposed based on the prediction of network bandwidth. Firstly, to get data information of the current network bandwidth available through online monitoring of the network in view of the networked control system with uncertain time-varying network bandwidth. Network prediction method is adopted to predict available network utilization of the next cycle online. Then the performance of the system are obtained by monitoring QoP(quality of performance) for the calculation of dynamic performance index IAE, and as a distribution standard of the available network resources prediction online, and realizes the system of network resources allocated dynamically. Calculate the sampling period by using the relationship between the sampling period and network bandwidth. Finally, validate the effectiveness of the system dynamic bandwidth allocation through simulation.