Research Of Network Control System Based On Time-Delay Compensation Algorithm

Network control system is the objective requirements for the complex control system and the remote control system, Compared with the traditional point to point control system, it has a wide range of applications prospects because of its easy to extend and maintain and ease to share of resources and so on, and thus become the hot spots of the research of control theory. Network is a double-edged sword, which might leads a lot of convenience also brings many problems, such as: network time delays, packet loss and network scheduling and so on. These problems make great difficulties to the design of network control system. In this paper, based on the analysis of network control system's characteristics and the research status of home and abroad, the issue of network time delays is the major research direction of my paper.For the issue of network time delays, this thesis made the work of two aspects as following:1. Time-delay estimation. First, a large number of measured time delay, based on the establishment of a network delay adaptive AR model, in both prediction accuracy and real-time basis, to determine the order of the model, and use the least mean square algorithm based on the new delay Value-line adjustment of the model weighting coefficient, by establishing a model to predict delay. Through the simulation of the time delay estimation algorithm in this paper have a good prediction of the delay effect.2. The time delay compensation. Based on the above time-delay estimation, for the issue of compensation for network delay, this paper presents two kinds of time-delay compensation algorithm as following:a. An improved fast generalized predictive control algorithm, to avoid the on-line solving Diophantine equations, reducing the computation time, and to better ensure the real time of network control system.b. Presents a different range falls under the time-delay Multi-model switching control strategy, when the time-delay is less than a sampling period, using the LQG optimal control algorithm because of better real-time; when the time-delay is greater than one sampling period, adopt to the improved fast generalized predictive control algorithms. And the simulation result show that the method has a good compensation effect on the time-delay and has a good real-time.