Robust hierarchical supervisory control of discrete-event systems

点击翻译按钮获取中文摘要

This research is concerned with the supervisory control of plants that can be modeled as discrete-event systems. A supervisory control system issues appropriate sequences of commands so that desired design specifications can be met. Examples of such sequences include the start-up and shutdown command sequences in a power or a chemical plant, or the command sequence issued to the propulsion system of a spacecraft during a maneuver. A typical plant consists of several subsystems. For such plants, supervisory control is usually designed following a hierarchical approach in which for each subsystem a local (low-level) controller is designed to enforce the specific design specifications of the subsystem, and a master (high-level) controller is designed to coordinate the operation of the subsystems. Compared with a supervisor with flat design, a hierarchical supervisory control system is more transparent and easier to design and maintain. In the proposed research, we intend to develop systematic methods for the design of hierarchical supervisory control systems. Computer code generation for supervisory control based on systematic methods is less prone to human error than manual code generation. Such systematic design methods are crucial for enhancing the autonomy and improving the reliability of controllers used in safety-critical and mission-critical systems. The applicant and his students have introduced a framework for designing robust hierarchical supervisory control systems. An important feature of this framework is that it can be used to design control systems that are robust with respect to modeling uncertainty (for example, due to failures). We plan to extend this framework so that if can guarantee deadlock and livelock free operation for the plant and to include timing information in issuing control commands. Furthermore, we intend to develop a decentralized version of this framework that is suitable for geographically distributed systems. Another area of application of our hierarchical setup that we would like to explore is hierarchical modeling of biological systems, specifically protein interactions in cells.