Some new architectures for fuzzy logic processing

Abstract

Fuzzy Logic has gained immense popularity in the last decade for use in several scientific and industrial applications. In this context, a substantial role is played by the computational structures used to perform fuzzy logic operations. For most real-time applications the software approach is too slow, requiring dedicated inference systems in hardware. Although many dedicated architectures for inferencing have been proposed in the literature, the need for finding new architectures that are more efficient in terms of higher processing speed and lower hardware complexity always exists. This research aims at the investigation of new fuzzy logic processing architectures. The major achievements of this research are the development of new dedicated hardware fuzzy inference system architectures that perform better in terms of both reduction in the time taken for a fuzzy inferencing cycle as well as in the hardware complexity. An investigation of the feasibility of multi microprocessor based cost-effective fuzzy logic processing systems has been carried out. It has been found that while speed improvements are possible with two or more processors, a system with two or three processing elements is most cost-effective. Also, by implementing fuzzy logic processing on the single and dual processor systems, it has been experimentally shown that dual processor systems can result in cost-effective improvement in processing for cases with two or more outputs.