BBPRM: A behavior-based probabilistic roadmap method

This paper focuses on the path planning problem. We offer an alternative to the probabilistic roadmap methods, from the perspective of modeling human or animal planning. In this context, hierarchies of representations are used to break down high-dimensional configuration spaces. We propose an approach for roadmap generation where low-level behaviors are used as articulations between level of the hierarchy. We also show how the obtained roadmap better represents low-level sensorimotor capabilities of the robot.