An optimization-based approach for fusing image-based trajectory generation with position-based visual servo control

Based on the potential for improving robustness to camera calibration effects, image-based visual servoing has been one of the most widely used control methods. However, some inherent technical problems associated with the use of a nonsquare image-Jacobian have been the subject of much discussion. In an effort to address these problems, a position-based visual servo controller is proposed in the presence of uncertain camera calibration and nonlinear radial distortion effects. Specifically, an optimization-based, online trajectory generator for the image features is fused with a position-based controller to move the kinematic system to an unknown desired task-space setpoint. A planar example based on a fixed camera configuration is used to illustrate the approach. The approach is then extended to the 6 degree-of-freedom case for the camera-in-hand configuration.