A hybrid approach for simultaneous obstacle avoidance and stabilization of dynamic bipedal walking using the aldebaran nao robot

In this paper, we propose a monocular vision system for autonomous obstacle avoidance and simultaneous stabilization of dynamic bipedal walking using the Aldebaran Nao robot. In particular, we address the case where erroneous bipedal locomotion causes the robot to drift away from a planned trajectory over time. To eliminate drifting, we use hybrid control patterns. This results in an efficient, self-correcting system that does not require to alter the robot's world representation. We confirm the efficiency and accuracy of the proposed system in a set of experiments.