The real-time implementation of fuzzy logic algorithms in embedded systems typically uses two approaches: employ fuzzy specific processing hardware or adapt standard embedded controllers to implement the fuzzy logic inference process. While high speed applications may require using the more sophisticated hardware, most embedded control applications do not have such processing speed demands, nor can they justify the added expense associated with the fuzzy enhanced processing engines. A review of embedded controller fuzzy logic implementations indicates a preference for 16-bit architectures; devoting significant processing resources to perform fuzzification, rule application, and defuzzification during real-time operation. While these approaches remain faithful to the foundations of fuzzy logic control, devoting processor resources to fuzzy specific tasks can limit a controller’s ability to handle peripheral tasks, such as man-machine I/O interface. This paper describes a simplified, hybrid approach suitable for standard 8-bit microcontrollers. The generic nature of the approach allows the methodology to be readily applicable to many single input, single output systems. This paper describes the hybrid fuzzy logic approach, which is placed in context using a proof-of-concept motor speed application. System performance data and notable limitations of the prototyped system are also described.
