The fields of computers and robotics have made impressive advancements within the past ten years. Compared with living things, the "evolution" of computers and robotics seems incomparably higher. For example, the evolution of the processing power and capability of computers has been amazingly high; so high in fact, as to double every year, or, I could even say, every half year. The field of robotics has also made steady progress thanks to the increased popularity and spread of industrial robots, which support the prosperity of the automobile industry. Honda's public announcement of its autonomous walking robot astonished the general public and attracted many peoples attention, making robotics a "fashionable" field as well.

The foundations of these brilliant advanced technologies is traditionally said to be the responsibility of "hard computing", which epitomizes and typifies thorough logical thinking. Theories based on "hard computing" are expected to evolve into more advanced matters through various breakthroughs, approaching the realization of the most ambitious engineering dreams. The news about Deep Blue, IBM's special "chess" computer that defeated chess champion Kasparov is a typical example of the realization of such dreams. On the other hand, the limits of "hard computing" are also becoming quite apparent: are computers, which execute computing at the nanosecond level, and human beings playing chess, driven by the same logic? Considering the information processing capabilities of neural cells, this is quite unlikely. Rather than invoke "hard computing" human beings should perhaps be regarded as performing "soft computing." Soft computing is a generic term for a collection of new approaches to information processing including neural networks, fuzzy sets and fuzzy logic, probabilistic theory and possibilistic algorithms, evidential reasoning, artificial life, etc. What unites these seemingly different computing approaches are the types of problems that they address, and how the solutions to these problems are obtained. Where the "hard computing" approach tailors the problem to the ideal solving paradigm, soft computing aims at providing theories general enough so that they can cope with more realistic problems. Consequently, soft computing takes an all inclusive view of the information processing task, exploiting, rather than eliminating features such as imprecision, inexactness, and redundancy of information. As expectations for these new approaches to information processing have been expanding, the importance of knowing the strategies of information processing in living things has been recognized.

From the viewpoint of engineering, living things are packages of wonders: the more technology develops, the more we are amazed at the remarkable mechanisms. This does not apply only to their brains. For instance, a little bird can fly using only the muscles in its body. Human beings have yet to invent an actuator both compact and light enough to mimic the efficiency of a birds muscles. It is also quite difficult to create a highly reliable organ such as the human heart, which works for more than 70 years without stopping. The most amazing organ is, of course, the brain, which is a combination of neural elements having simple computing capabilities, operated by self-organization based on local information transmissions. By these means, the brain achieves vigorous distributed expressions and executes flexible, efficient, and virtually trouble free computing. Existing engineering methods have failed at simulating or duplicating these wondrous processes.