Learning Systems and Intelligent Robots

This book contains the Proceedings of the S~cond U. S. -Japan Seminar on Learning Control and Intelligent Control. The seminar, held at Gainesville, Florida, from October 22 to 26, 1973, was sponsored by the U. S. -Japan Cooperative Science Program, jointly supported by the National Science Foundation and the Japan Society for the Promotion of Science. The full texts of the twenty-one presented papers are included. The papers cover a variety of topics related to learning control and intelligent control, ranging from pattern recognition to system identification, from learning control to intelligent robots. During the past decade, there has been a considerable increase of interest in problems of machine learning, systems which exhibit learning behavior. In designing a system, if the a priori infor­ mation required is unknown or incompletely known, one approach is to design a system which is capable of learning the unknown infor­ mation during its operation. The learned information will then be used to improve the system's performance. This approach has been used in the design of pattern recognition systems, automatic control systems and system identification algorithms. If we naturally extend our goal to the design of systems which will behave more and more intelligently, learning systems research is only a preliminary step towards a general concept of integrated intelligent systems. One example of this class of systems is the intelligent robot, which integrates pattern recognition. learning and problem-solving into one intelligent system. Front Matter....Pages i-ix The Concept of a Linguistic Variable and its Application to Approximate Reasoning....Pages 1-10 Fundamental Concepts and Social Consequences of Artificial Intelligence....Pages 11-31 Biorobots for Simulation Studies of Learning and Intelligent Controls....Pages 33-46 A Mathematical Neuron Model which has a Staircaselike Response Characteristic....Pages 47-91 Performance Aspects of Stochastic Nonlinear System Classification by Pattern Recognition Methods....Pages 93-113 Algorithmic Techniques for Modeling Nonlinear Functions....Pages 115-144 A Survey of Heuristic Search Method of Multimodal Optimum Point....Pages 145-169 Basic Search Patterns in Heuristic Search....Pages 171-190 Multi-Modal System Identifications by a Learning Procedure....Pages 191-210 Learning Dual Control under Complete State Information....Pages 211-231 On a Class of Variable-Structure Systems....Pages 233-248 A Method of Learning Control Varying Search Domain by Fuzzy Automata....Pages 249-262 Adaptive Computer Aiding in Dynamic Decision Processes....Pages 263-271 Optimal Learning Recognizer for Unknown Signal Sets in a Channel with Feedback Link....Pages 273-294 Computational Algorithms for Interactive Pattern Recognition....Pages 295-316 A Methodology for Interactive Systems....Pages 317-324 Automatic Recognition of Complex Three-Dimensional Objects from Optical Images....Pages 325-341 Eyes of the Wabot....Pages 343-364 The “Rubber-Mask” Technique-I, Pattern Measurement and Analysis....Pages 365-400 The “Rubber-Mask” Technique-II, Pattern Storage and Recognition....Pages 401-421 Learning Texture Information from Singular Photographs and its Application in Digital Image Classification....Pages 423-435 A Theory of Character Recognition by Pattern Matching Method....Pages 437-450 Back Matter....Pages 451-452 This book contains the Proceedings of the S̃cond U.S.-Japan Seminar on Learning Control and Intelligent Control. The seminar, held at Gainesville, Florida, from October 22 to 26, 1973, was sponsored by the U.S.-Japan Cooperative Science Program, jointly supported by the National Science Foundation and the Japan Society for the Promotion of Science. The full texts of the twenty-one presented papers are included. The papers cover a variety of topics related to learning control and intelligent control, ranging from pattern recognition to system identification, from learning control to intelligent robots. During the past decade, there has been a considerable increase of interest in problems of machine learning, systems which exhibit learning behavior. In designing a system, if the a priori inforƯ mation required is unknown or incompletely known, one approach is to design a system which is capable of learning the unknown inforƯ mation during its operation. The learned information will then be used to improve the system's performance. This approach has been used in the design of pattern recognition systems, automatic control systems and system identification algorithms. If we naturally extend our goal to the design of systems which will behave more and more intelligently, learning systems research is only a preliminary step towards a general concept of integrated intelligent systems. One example of this class of systems is the intelligent robot, which integrates pattern recognition. learning and problem-solving into one intelligent system