Networked Control Systems: Algorithms and Experiments

Abstract

In this chapter we address the problem of implementing control systems using general purpose communication networks to transmit plant state information and control signals. The main idea is to expose the issues that have to be considered when implementing teleoperation and telepresence applications exploiting already deployed communication infrastructure. Some of these applications may be implemented with local controllers and remote supervisory systems, where the remote client sends set points according to the process status received, but the control loops are closed locally. Other applications however, might require gathering information from different and geographically distant agents or sensors. In such a case, loops cannot be closed locally, and the state and control signals must travel across the networks. A general purpose communication network will however introduce issues such as propagation time-delays and loss of information. Therefore, the control algorithms must now account for these issues, and they should be robust enough to guarantee a certain level of performance. We develop in this section a series of experiments to identify the issues induced by a general purpose communication network, with specific emphasis on wireless networks. We use standard operating systems and industrial hardware for data acquisition. Then, we propose compensation alternatives to cope with these issues.

The research of all authors is partially supported by NSF-0233205.