Feature based robotic visual tracking of 3-D translational motion

The authors address the problem of robotic visual tracking of a target that moves in 3-D with translational motion. The camera is mounted on the end-effector of the robotic manipulator. Use is made of simple SISO (single-input single-output) adaptive controllers in conjunction with a Cartesian robotic controller for successful tracking. The scheme does not need explicit computation of the target's relative depth with respect to the camera frame. In addition, the relative motion of the target with respect to the camera is measured by using a cross-correlation technique called sum-of-squared differences optical flow. Experimental and simulation results are presented to show the efficacy of the proposed approach.