Network Control System Delay Analysis And Its Controller Design

Network control system (NCS) is a comprehensive interdiscipline which developed rapidly with control technology, computer network technology and communication technology in the past ten years, NCS got great attention by researchers around the world. Network control system has many advantages when contrast with traditional control system, nevertheless, because closed-loop network is introduced to system, in order to make the whole control system can still achieve better performance, there are still many problems to be studied and solved. Among them, the network time delay is the main factors in entire network control system which affect control performance, if the network delay is big enough, the whole control system will be not stable. In order to minimize or eliminate the effects of network delay to the control system, based on the researches of network control system theory study, this paper studies some problems of the control algorithm and the simulation tools, the main research contents are as follows:Smith estimated compensator was introduced in order to offset the influence of fixed network delay when the network delayed very little. The fuzzy adjustment algorithm is introduced to NCS, according to the error signals and error signal's rate of input, the adjusting rules of fuzzy PID controller parameters are set to adjust the PID parameters constantly. Using the MATLAB simulation platform, through the DC motor of the simulation experiments show, the estimate compensation based on Smith fuzzy PID control than fuzzy PID control has better control effect.For most network delay more than one sampling period, the control algorithm based on the model state prediction is designed. The state observer is set in a sensor, the state observing value is calculated and transmitted to the controller through network, the controller calculated the controlled object future time according to controlled object model, by state feedback, the control information is transmitted to actuators through network, data buffers is set in actuators end, actuator according to control information sampling serial number and current sensor sampling serial number difference to choose appropriate control information in the role of controller. Simulation results show that the control algorithm is effective.Aiming at network random delay, the controller is designed by using neural network to predict future delay and combined with internal model control strategy in this paper. A mature BP neural network algorithm is applied to delay predicted, the BP neural network is trained using two delay values in the internal model controller, and then next delay value is predicted, the nicer control effect is obtained by adjusting the filters' filtering coefficients. Therefore, selecting of the network delay prediction precision and filters' filtering coefficients become the keys of network control system design.