A systolic array circuit for the fuzzy inference process

The fuzzy inference process is mostly used in control applications today. Investigations show that for possible future application cases, for instance fuzzy expert systems and multi-stage controllers, sup-min composition in combination with Godel implication will achieve better results than the match factor method which is generally used in fuzzy hard- and software for cost reasons. This permits the chaining of inference units for more complex fuzzy systems. However, the calculation expenses for sup-min composition with Godel implication are high, therefore the use of a processor array becomes necessary. After a short discussion of the algorithm the prototype of a Godel implication based, cascadable array circuit for fuzzy inputs and outputs will be presented.<
>