Position predictive control of an anthropomorphic robotic arm using a time-of-flight camera

This paper introduces a predictive control strategy for visual servoing an anthropomorphic robotic arm in an object tracking task. Robot workspace constraints and the target location are addressed through a time-of-flight camera. We present a methodology to process the vision information to explicitly identify and remove obstacles and objects from the point cloud. Later, this information is used by a predictive-control strategy, achieving convergence during the robotic arm motion and avoiding obstacles that are included in the cost function as constraints. The proposal validation was successfully tested on a real robotic set-up.