Delay Compensation Method Of Missile Networked Control Systems

The problems of the traditional point-to-point control systems structure areincreasingly prominent, for example, complex wiring, difficult maintenance, poorscalability, high cost, which can’t meet the continuously improving the requirement ofperformance of missile control system. In order to overcome these defects, more andmore the network control technology is used in missile. Due to the limitation of busnetwork bandwidth and network protocol, the information generates delay. By thecomparative study on the features of the1553B bus, CAN bus and Ethernet, this paperchooses the CAN bus as the control network of bus network control system to study,analyzes the generation and composition of network delay on the missile networkcontrol, establishes the delay model of bus network control system.After comparative analysis of common delay compensation scheme, consideringthe influence of model simplification to the actual control network, this topicestablishes the bus network control system on the delay characteristics of missile busnetwork, and design a delay compensation strategy using RBF neural networkpredictive control based on predicted deviation, to achieve real-time compensation ofnetwork delay, then MATLAB is used to accomplish calculation and simulation.Compared with the method of PID control, the method of RBF neural networkpredictive control based on predicted deviation, can effectively improve the effects ofnetwork delay, reduce the overshoot, and shorten the setting time.In order to overcome the limitations of delay module in the MATLAB simulation,after delving into the influence of the network delay on missile bus network controlsystem, this topic takes the TRUETIME simulation toolbox based onMATLAB/SIMULINK, reproduce the network environment of bus really to discussthe effect of network delay on performance of missile system. Use TRUETIME is usedto realize simulation of missile bus network control system under different time delayand scheduling policy, and there is a comparative study between method of RBF neural network predictive control and method of PID control in the networkenvironment. By changing the different way of bus network transmission, verify theperformance of1553B bus, CAN bus and industrial Ethernet network on controlperformance and information transmission performance. Results show that the RBFneural network predictive control method can more effectively improve the effects ofnetwork delay.