Compiling Fuzzy Logic Control Rules to Hardware Implementations
This work addresses the need for efficient real-time control in complex systems, though it appears incremental as it builds on existing hardware methods for fuzzy inferencing.
The paper tackles the problem of implementing fuzzy logic control for real-time applications by developing a programming environment that translates fuzzy control rules into hardware, enabling faster processing to meet stringent speed requirements.
A major aspect of human reasoning involves the use of approximations. Particularly in situations where the decision-making process is under stringent time constraints, decisions are based largely on approximate, qualitative assessments of the situations. Our work is concerned with the application of approximate reasoning to real-time control. Because of the stringent processing speed requirements in such applications, hardware implementations of fuzzy logic inferencing are being pursued. We describe a programming environment for translating fuzzy control rules into hardware realizations. Two methods of hardware realizations are possible. The First is based on a special purpose chip for fuzzy inferencing. The second is based on a simple memory chip. The ability to directly translate a set of decision rules into hardware implementations is expected to make fuzzy control an increasingly practical approach to the control of complex systems.