Hybrid intelligent techniques into autonomous control systems

Abstract

Intelligent systems which perform tasks autonomously are required in many fields such as factories, nuclear power plants, space robots, and medical fields. Such systems must have many kinds of abilities such as sensing, perception, knowledge acquisition, decision making, learning, inference and acting functions, and they have to integrate these abilities so as to adapt to external unknown environments and given tasks. Recently, computational intelligence including neural networks, fuzzy systems and genetic algorithms has been discussed for realising human intelligence. The paper presents the major advanced techniques of building complex hybrid systems for intelligent control. An experimental environment which integrates these paradigms is proposed. Its main function is to assist the human designer in choosing the best strategy of developing an intelligent control system by making an "optimal" combination of these above mentioned four paradigms.