ABSTRACT
Because of the human limitations to supervise every simple activity and task in uncertain environments, it is highly desirable to devise machines that can perform the same tasks with minimal interaction with a human operator. These machines are called intelligent machines (IM). This chapter discusses a general theory of IMs and presents the theory and mathematics for the design of intelligent robotic systems. It describes very briefly systems developed for particular applications. In contrast to knowledge-based (expert) systems that use heuristic methods in problem solving and job execution, hierarchically intelligent control applies a mathematical approach for the same tasks. It is based on a three-level interactive probabilistic model that utilizes the associated entropy as a measure of success of the job and accommodates fast and reliable operation of the functions of the IM. IMs are designed to operate in structured or unstructured uncertain environments with minimal supervision and interaction with a human operator.
