I. Introduction
An advancing trend is to substitute specialized industrial networks with a general network such as Ethernet to control applications in distributed control, industrial electronics, and factory automation areas remotely over an Internet Protocol (IP) network [1]–[3]. A protocol like Ethernet has several advantages due to its well-developed infrastructure, widespread usage, and affordability, for IP connection. Of course, a controller can be connected to a plant to form a local closed-loop control system and the status of the system can be monitored via an IP network as well as simple on-off control. However, this configuration does not have good interaction with an operator or an automatic controller located on a remote side. When an anomaly occurs, the remote side can only notice it from reported messages and may not be able to respond in time.
Overall distributed networked control system over an IP network.
Researchers have proposed a scheme to have a controller connected to a plant via a network as a closed-loop networked control system [1] so that the remote controller could interact with the plant via the network more quickly. Several methodologies on this topic have been developed to handle the time-varying network delay effects. These methodologies are based on various techniques such as state augmentation [4], optimal stochastic control [5], nonlinear control [6], robust control [7], buffering and prediction [8], [9], Smith predictor [10], event-based control [11], sampling time scheduling [12], [13], optimal gain scheduling [14], scattering transformation [15], and fuzzy logic [16], [17]. Many of these techniques require a completely redesigned controller to handle network delays. Practical controllers that exist in industrial applications such as a proportional-integral (PI) controller have to be replaced. This replacement is time-consuming and costly.