Mapping human´s intelligence in the real domain-the theory and its practice

Originally, fuzzy inference was introduced to help translating vague, imprecise human reasoning into numerical algorithms so that computers can handle. We argue, however, that in most cases such mapping of human´s intelligence can be done completely in the real domain, fuzzification and defuzzification being merely transformations of data that could also, and sometimes more naturally, be handled untransformed. This offers the following advantages over the conventional fuzzy approach: 1. Simpler hardware implementation. 2. More transparent input-output relationship for ease of analysis and design