Extraction and transfer of fuzzy control rules for sensor-based robotic operations

Human beings can easily describe their control behaviours with IF-THEN rules, which can be transferred to similar tasks with only slight local changes. However, standard techniques for function approximation like neural networks or associative memories are unable to work with rules. We introduce a method for extracting and importing rules from and to a B-spline fuzzy controller. Rule importation is used to initialise a B-spline fuzzy controller with a priori knowledge to reduce the learning time and to improve the poor generalisation capability. Through "learning-by-doing" and "learning-from-others", set of rules for a two-arm cooperation task can be generated and further transferred to a robot screwing operation and a three-arm cooperation. The successful experiments show how rule-based knowledge can be used for skill transfer in analogue tasks to help significantly accelerating a learning process. (C) 2002 Elsevier Science B.V. All rights reserved.