A behavior-based implicit planning method is proposed through the design of a middle-size autonomous robot soccer system (MARSS). With basic goal-driven behaviors, the MARSS reactively selects suitable basic behavior according to different situation. The reactive executions of basic behavior sequences implicitly plan two primary behavior chains. By the robot's dynamical interactions with the environments, interleaving of basic behaviors spontaneously exhibits effective emergent behaviors to deal with some difficult situations. These emergent behaviors make the system simple, robust and competitive.
展开▼