Analysis, design, modeling, and control of networked control systems

A major trend in modern industrial and commercial systems is to integrate computing, communication, and control into different levels of machine/factory operations and information processes. These distributed control systems are called Networked Control Systems (NCSs), and have sensors, actuators, and controllers interconnected by communication networks. The change of communication architecture from point-to-point to this common-bus approach, however, introduces different forms of time-delay uncertainty in the closed-loop system dynamics. These time delays come from the time sharing of the communication medium as well as the computation time required for physical signal coding and communication processing. Moreover, the time delays in a control application can degrade a system's performance and even cause system instability.; To understand the time delays in an NCS due to networked communication, the medium access control mechanism of three candidate control networks, ControlNet, DeviceNet, and Ethernet, is discussed in detail, and the key performance parameters are identified. The time-delay characterizations of the network transmission and device processing are theoretically formulated and verified by simulation and experimental analysis. Simulation and experimental studies further validate the impact of network-induced delays on control performance.; In digital control, control performance mainly depends on the control system bandwidth and sampling rate. In networked control, the control performance is also affected by network parameters. Based on the characterization of the network-induced delays, we propose an NCS design chart and a systematic design consideration algorithm. The NCS design chart and design algorithm utilize both network and control parameters to reveal the existence and location of the performance degradation points which limit the operating range of sampling periods.; After the NCS design chart is identified, two methods for performance improvement are presented in this dissertation. The first method tunes the network system parameters such as the prioritization of different subsystems or nodes and the choice of different types of network protocols. The second method uses an exact model of the NCS with delays to design an optimal (LQR) controller. Simulation studies demonstrate the performance improvement achieved using these methods.; In order to verify and illustrate the theoretical results, we also develop a Networked Control Systems Toolkit, based on Matlab/Simulink™. The NCS toolkit is used to model processing time delay, simulate network traffic and control performance, and design delayed optimal controllers. A DeviceNet network is also implemented on a three-axis machining tool at ERC/RmS, the University of Michigan, to validate the theoretical and simulation results.