Gaussian-basis-function neural network control system with network-induced delays

The current development of communication technology makes it possible for control engineers to implement networked architectures for distributed control systems. However, the data transfer over communication networks inevitably adds some non-deterministic delays in the closed-loop systems. The analysis and design of this type of systems has attracted a great deal of research interest over the world. In this paper, a laboratory system is used to study the issue of non-deterministic delays in a networked control system. The study is based on a physical platform where three computers, functioned as a digital model of a plant, a system identifier and a learning controller respectively, are communicated through an Ethernet. In this system, identification and learning control are based on Gaussian-function neural network algorithms. Some network-induced delay in our three-computer-Ethernet platform is measured and its effect on the control system is analysed. To reduce the detrimental effect caused by the delays, a scheduling approach, tested on our laboratory system with favourable results, is presented in this paper. Due to its simplicity and effectiveness, it is believed that it may be possible to apply the proposed approach to some practical networked control systems.