Prediction Of Random Time Delay In Networked Control System Based OnAuto Regression Analysis

With the rapid development of the Internet, the combination of networked and control system has gradually become a hot research topic. Among them, the research of networked control system (NCS) has become an advanced subject in the field of automation. Networked control system is a distributed real-time feedback control system which through the date exchange between sensors, controllers and actuators. The system has the advantages of simple installation, reliable system, remote control and so on. However, after the introduction of networked in control system, due to the limited network bandwidth, the control system will inevitably lead to the network induced delay in the network transmission. Due to the network load, node competition, network congestion and other factors, the network delay presents the randomness. Network delay becomes the main reasons for the networked control system performance degradation and even instability, how to find an effective network delay prediction method plays an important role in networked control system modeling and control researching.In the current network time delay prediction of networked control system, there exists a non-stationary series analysis effect is poor, multiple step prediction error is large, the lack of software implementation of the system and so on. In this background, in order to compensate for the network delay, this paper based on network delay prediction research of networked control system, propose Auto regressive Integrated Moving Average(ARIMA) model, effectively combine the data preconditioning, difference stationary process, minimum mean square error algorithms, to predict the network time delay of networked control system. The specific content of the following aspects:(i)By using OPNET software to build network topology structure, set various parameters of the networked control system, after running the network to measurement network delay, obtained a large number of random delay data of the networked control system.(ii)To test the hypothesis of random time delay in networked control system, if it does not satisfy the conditions of stability, the random time delay sequence is processed and tested until it meets the condition of stability.(iii)The prediction algorithm based on ARIMA model is proposed for the system research of various forecasting methods.