ABSTRACT
Skill-based expert system controlling robot operations is an excellent example of a nonnumerical modeling approach. This chapter describes the skill-based expert system for the control of industrial manipulators which was designed for robots with dexterous hands. It deals with the overall control aspects of such a manipulator. The chapter considers general issues related to the design of smart hands. A general methodology for teaching motor skills to the machine is not available. The methods of knowledge capturing described are the product of some basic neurophysiologic considerations and heuristic approaches. The relevance of synergy phenomena to skill-based expert systems can be explained by several examples. An interesting application of the scene simplification principle for knowledge capturing pertinent to manipulation tasks is presented when describing the skill-based expert system for the control of industrial manipulators with multifingered hands.
