Projection homography based uncalibrated visual servoing of wheeled mobile robots

Abstract

In this paper, a novel visual servo regulation strategy, which is based on the projection homography matrix, is presented for a wheeled mobile robot in the presence of unknown camera intrinsic parameters. Specifically, the derivation of the projection homography matrix is described and then estimated without being restricted by the scale factor. Subsequently, to deal with the unknown depth information and unknown camera intrinsic parameters, an adaptive controller is carefully developed by utilizing the entities of the projection homography matrix to drive the mobile robot to the desired position. Using Lyapunov techniques, rigorous stability analysis is conducted. Afterwards, a controller for the angular velocity is further designed to regulate the orientation error of the mobile robot. The proposed method does not need any complex pose estimation algorithms, and it thus avoids the confusion made by the two solutions provided by homography decomposition methods. The performance of the proposed uncalibrated visual servoing strategy is further validated by simulation results.