Visual servoing based positioning and object tracking on humanoid robot

Abstract

In this paper, we present a reactive vision based autonomous behavior on a humanoid robot. This perception action scheme enables the robot to walk toward a desired position or to track a given target and grasp it. The proposed solution uses visual servoing as a reactive sensor based technique to address the problem of perception, thus to ensure reachability of the target object by generating appropriate motion commands. An omnidirectional velocity tracking gait generator is considered to solve the locomotion and balance problem. It is based on the framework developed by Herdt et al. (Advanced Robotics 24(5–6): 719–737, 2010; 2010 IEEE/RSJ International Conference on Intelligent Robots and Systems, pp. 190–195, 2010), which has been extended for instantaneous rotational velocity in this work. The redundancy problem due to the high number of degrees of freedom associated with the mobility in operational space is handled by a task sequencing technique. This allows the improvement of reachability and manipulability while accounting for joints limits and obstacles avoidance. Finally, simulations on the humanoid robot NAO are shown to evaluate the effectiveness of the proposed scheme.