The Center for Automation and Intelligent Systems Research
has been affiliated with the Cleveland Advanced Manufacturing Program
since its inception in 1984. Caisr's mission is to promote excellence
and growth in the following three areas.
In 1952, Professor Donald Eckman of the Case Institute of Technology had the vision to predict the tremendous impact of digital computers applied to the control of industrial processes. In pursuit of this vision, he founded the Systems Research Center, an interdepartmental center, one of whose divisions was the control of industrial Systems Program, a unique collaborative research effort between industry and Case faculty. The Center for Automation and Intelligent Systems Research is the descendent of Eckman's System Research Center and so, it is natural that the Control of Industrial Systems Program be a part of CAISR.
As Eckman predicted, availability of relatively low cost computing power has vastly broadened the domain of what is technologically and economically feasible to achieve in the control of industrial systems. These factors have motivated the development of theory, approaches and design methodologies which exploit new opportunities for implementing sophisticated control and information processing algorithms, for distributing control functions, for adding "intelligence" and for achieving system integration based on hierarchical control.
The Center is in the forefront of research on artificial neural networks. Systems have been and are being developed using this technology to address practical problems in control and in "modeling" complex industrial processes. Even though this technology is relatively new, currently available software has tremendous power in "discovering" features of systems that have heretofore resisted traditional methods of modeling and analysis.
This is due to the fact that researchers have made great strides in applying computers to model artificial neural networks. This technique modeled after biological systems, has demonstrated the ability to "discern" features and "learn" interrelationships in terms that add new insight into the overall problem. An additional benefit of these networks is that their performance degrades gracefully. If the network is damaged or if bad data enters the system, the network supplies an answer close to the desired result as opposed to "crashing" as is the case with other computer based approaches
Among the most valuable resource in any company is the accumulated expertise of its employees. This knowledge, obtained at considerable investment of both time and money, forms the basis of a company's current operations and the starting point of future growth. In many cases this information resides in the minds of key personnel, much of it undocumented, in formal records of the company. The use of expert systems is an excellent means of providing ready access to this knowledge. Once it is captured in the system, the ability to make consistent descisions in the subject area is available to others within the organization.
In addition to supporting descision making, the expert system serves as a formal repository for existing knowledge and provides a means of training new employees. Key personnel are freed from the need to address repetitive problems or training exercises and are allowed to concentrate on the truly novel situatiions that arise during the course of an activity.
Expert Systems are a major component of a branch of computer science called "artificial intelligence". An Ai system addresses problems by employing the characteristics normally associated with the human mind -- the ability to use "learning", "reasoning" and "language comprehension" in the solution of problems.
Research in the Mechatronics Research Laboratory is concerned with enhanced performance of computer-controlled machinery. Performance enhancements include increasing speed through design and control, providing new capabilities such as cooperative motion among multiple machines and creating control structures which promote effective integration with artificial intelligence. The overall objective iis to increase intelligent machine autonomy and increase performance.
The center hosts the Network Research Lab whose main objectives are to pursue basic research in networking and corresponding applications to remote robotics and control. The major thrusts in this area are to increase the scalability of the current Internet infrastructure through multicast methods and the related date management components, to examine the behavior of IP networks with Quality-of-Service provisioning, to devise middleware for pervasive computing, and to establish an agent-based franework for remote robotic manipulation.
Contact the Webmaster at root@r2d2.eeap.cwru.edu
or contact CAISR director at smp2@po.cwru.edu
Back to CAISR home page