The identifier-based relative position estimation for leader-follower robotic system

A monocular vision measurement approach is presented for leader-follower robotic system to estimate the position of the leader robot relative to the follower robot. The color identifier with green and red blocks is attached on the back of the leader robot, and it may be captured by the camera mounted on the follower robot. A coarse-to-fine searching method is given to extract the center point of the identifier. The parameters of the camera with the combination of the intrinsic and extrinsic ones are calibrated via least square method with given relative positions and the corresponding image coordinates of the identifier's center. The parameters are further optimized with Levenberg-Marquardt method. Then the position of the leader robot relative to the follower robot can be estimated only with single feature point. The experimental results verify the validity of the given approach.