Contents I. Introduction
THE study of control systems in which sensor data are fed back using a data network has recently received much attention. The use of a data network in the feedback path has several advantages such as reconfigurability, low installation cost, and easy maintenance; it is also well suited for large geographically distributed systems. There are, nevertheless, some shortcomings to the use of digital networks in control systems such as the fact that data networks operate in a discrete fashion delivering information only at specific instants in time. This means that the controller cannot have access to the plant output at all times. Moreover, increasing the availability of the output information to the controller means increasing the bandwidth needed by the control system. Even for discrete plants where the controller acts only at specific instants of time, if the sensor data is sent through the network at each sampling time and the sampling rate is low, the bandwidth required might still be large.
