An Intelligent Scalable Control Architecture For Distributed Nonlinear Systems

This project will develop and implement a learning controller architecture for control of complex distributed nonlinear systems with multiple localized equipment controllers and a variety of sensor feedback. The fundamental philosophy of this approach is to optimize control at a local level through the use of neural networks and optimal control techniques such as dynamic programming, and to utilize the longer term planning and optimization capabilities of optimal control to coordinate the local controllers and achieve global system optimzation. This intelligent control architecture will be implemented on an advanced scalable distributed control platform developed by Honeywell, the Advanced Equipment Controller System. This work will build upon numerous previous research and development projects in intelligent and distributed control, both by NeuroDyne Inc. and Honeywell, under funding from DARPA, NSF, NAVY, NASA, Air Force, internal R&D, and others. This will result in a flexible, intelligent, scalable, modular, adapative distributed control platform which may be easily applied to a wide variety of systems and processes. ANTICIPATED BENEFITS: The potential commercial applications of this technology are as wide as the applications of industrial control itself. A proven intelligent control architecture will be implemented on an advanced Honeywell scalable controller hardware platform to provide self-configuring, self-optimizing process control which is easily adaptable to a wide variety of industrial processes in aerospace, microelectronics fabrication, chemical process control, and numerous other areas.