Fast fuzzy signal and image processing hardware

The paper presents the development of fast fuzzy logic based hardware for various applications such as controllers for very fast processes, real-time image processing and pattern recognition. It is based on the fired-rules-hyper-cube (FRHC) concept, characterized by extremely simple way of the fuzzy inference in a layered parallel architecture. The processing time slightly depends on the number of inputs of the fuzzy system and does not depend on the number of rules and fuzzy partitioning of all variables. Most important is the inherent high speed of processing because of the parallelism and pipelining, implemented in all layers.