CET4xx - Neural & Fuzzy Logic Systems
J. Sumey 29-Oct-98
This upper-level course explores areas of alternative computer engineering methodologies and provides roughly equal coverage of neural networks and fuzzy logic systems. Neural-based logic systems attempt to solve problems by approximating the operation of the human nervous system instead of the procedural methods typically taught in computer science. Neural networks are particularly useful in pattern matching type applications. Fuzzy logic systems also seek to solve problems, such as the inverted pendulum problem, from a more abstract, non-procedural approach whereby results are determined by “strength of truth” of membership functions. The components of a fuzzy system including input fuzzification, inference engine and rule set, and output defuzzification are examined. Both neural and fuzzy methodologies are being applied more frequently as computers are being used to solve larger and more complicated problems as well as in many embedded systems. Being a laboratory course, the student will complete a number of hands-on projects involving software and/or microprocessor-based systems that demonstrate practical, real-world applications of neural and fuzzy logic systems.