TABLE OF CONTENTS
MOST GENERIC FORM OF AI FOR EMULATION OF HUMAN THINKING
NEUROCIMPUTATION IS INSPIRED BY BIOLOGICAL NEURAL NETWORK OF HUMAN BEING
BASICALLY INPUT-OUTPUT NONLINEAR MAPPING PHENOMENA LIKE FUZZY SYSTEM
MASSIVE HIGH SPEED PARALLEL COMPUTATION WITH FAULT-TOLERANCE AND NOISE FILTERING CAPABILITY
KNOWLEDGE IS ACQUIRED BY LEARNING (OR TRAININING) THROUGH EXAMPLES OF INPUT-OUTPUT DATA SETS
PROPERTIES OF
PATTERN CLASSIFICATION AND RECOGNITION
FUNCTION APPROXIMATION
ASSOCIATIVE MEMORY
TYPICAL APPLICATIONS:
CONTROL AND ESTIMATION IN POWER ELECTRONIC SYSTEMS
GENERAL INDUSTRIAL PROCESS CONTROL
ROBOT VISION
ON-LINE DIAGNOSTICS, ETC.
Among all the AI techniques, artificial neural network (ANN) or neural network (NNW) is the most important discipline, and its potential impact on power electronics area is tremendous. The technology has a long history, but its development was camouflaged by the glamorous evolution of modern digital computers. From the early nineties, the momentum of its R & D and applications has surged dramatically. As mentioned before, neurocomputer attempts to mimic the capability of biological nervous system, but obviously its performance is far inferior. Like fuzzy system, NNW basically performs input-output mapping which can be static or dynamic. The result is pattern recognition, pattern classification, function approximation and associative memory properties which will be discussed later. One important feature of NNW is that it normally requires supervised training (or learning) by input-output example data sets unlike conventional programming of digital computer. NNW is a vast subject [1, 2]. We will discuss its principles and applications in power electronics and motor drives in this chapter.
