A rule-based decision subsystem design approach for intelligent robot

摘要

A rule-based forward reference approach was proposed for the decision subsystem of the intelligent robot to avoid the complex decision logic and difficulties in improving the performance of the heuristic direct control solutions and to decrease the response time of machine-learning based intelligent approaches. The performance of this rule-based approach can be enhanced by modifying and improving the knowledge base only due to the characteristic that knowledge and control logic are separate and independent of each other in knowledge based system. Moreover, “primitive action”, “priority of parallel conclusion”, “parallel search method” and “simple conflict resolution method” were also introduced to shorten the system's response time. The successful prototype application in robot soccer verified its effectiveness.