Time Delay Compensation For Network Control Systems And Its Application

The network control system is a real-time feedback control system, which contains network in some positions of the traditional control system. It can implement the functions of sharing resources or coordinating operations between the users who are not in the same place but in a defined geographic area. The sensor, the controller and actuators are connected by a common communication medium, so in the process of transmission for some information, the information can collide each other and the information sometimes can be resent and so on. Delay generated in this process is called network-induced time delay. This is one of the main problems of the network control system. The existence of the time delay can influence the performance of the network control system:such as the range of stable, the time of settling and rising, in some serious cases the system even can not be maintained in a stable state. Therefore taking some certain methods to compensate the delay of the network control system is very important to the the real-time network control system, and it is very theoretical and practical. This paper mainly studies the compensation for the network control system, and this paper mainly concludes the following contents.First for the network control system with short time delay, we can establish a piece-wise function according to its control input which was the influenced by the delay. Then we can get its model when apply the function to the control system. A state observer was constructed to actively compensate the network transmission delay. Later we rebuild an augmented state, then deduce a stable condition for the network control system with short time delay. At last we can get the compensation controllers.Secondly for the network control system with long time delay, first we consider the case where there is no packet loss. A control prediction generator is designed to compensate the time delay in the network control system. By selecting a reasonable Lyapunov function, we can solve the problem of the system's stability and the network delay compensation controllers; then we consider the case where there exists packet loss:we still use the control prediction generator to achieve the compensation for the time delay. The problem of packet loss can be modeling as a Bernoulli random sequence, therefore, the model of networked control systems established will become a randomized control system model. Then we will study the stability of the system and the expression of the network delay compensation controllers.At last, a mobile robot was selected to verify the correctness of the theory. First we can construct its kinematics model according to its structure, then get its model with delay according to its analysis. We can obtain its controllers according to the theories above. Through comparing the differences between the case with compensation and without compensation by using simulation diagram, we can prove the necessitv of compensation and the correctness of the theory.