Computer-Aided Control Engineering (CACE) Tools: AI Techniques and Applications

The future of intelligent control systems depends upon the extent to which Artificial Intelligence (AI) technology can help control engineers deliver practical solutions to difficult control engineering problems. Conventional control design approaches have achieved notable successes in the design and implementation of robust, adaptive controllers for systems with well-defined mathematical models. However, conventional approaches have had difficulty supporting engineers in the design and implementation of control systems when an accurate mathematical model is not available. Also, verification that computer-controlled systems perform to specifications, validation of the specifications, higher-level control, operator decision aids, system diagnosis, operator alerting, and reconfiguration of systems which experience large changes over time or potentially catastrophic failures are significant challenges to control science and engineering. It is in these difficult areas where the AI technologies of knowledge representation, learning, search, diagnosis, planning, and decision are being used to aid control engineers. Algorithms for computer-controlled systems and software tools to help implement these algorithms have been a subject of research and commercialization for decades. Computer-Aided Control Engineering (CACE) tools have achieved a degree of success in the past decade based on their ability to assist in the control system design and implementation process. Specialized tools have been made available for system identification, system simulation, controller design and controller implementation. Recently, efforts have been made to build integrated CACE environments. Also, some current research is aimed at increasing the utility of available systems by creating a mathematical basis and a software architecture for efficiently describing complex systems and using these as a means of achieving a higher level of integration of the diverse tools already available. A recent Workshop on Software Tools for Distributed Intelligent Control Systems was sponsored by the U.S. Army and The Defense Advanced Research Projects Agency (DARPA). This paper will describe the results of the workshop and subsequent efforts to use these results to shape a DARPA software development project. The first section of the paper provides a brief review of the current applications of AI in the design and implementation of control systems. The second section discusses areas where AI can be applied in the near term to help solve challenges in the implementation of computer-controlled systems. The third section gives an overview of the development of CACE tools. The fourth section provides a review of the Army/DARPA workshop and the last section discusses the use of the results of the workshop.