Evaluate Control Laws Related To Interaction Matrix For Image-Based Visual Servoing

In this paper, the methods for approximating the interaction matrix used in image-based visual servoing are analyzed. The five approximation methods are reviewed and evaluated based on the ability to solve the issues related to the interaction-matrix approximations such as the camera retreat problem, out of camera's field of view, singularity, stability and local minimum. In general, behavior of the methods is different although some methods have similar behavior in a few special case and each method has its own advantages and disadvantages Therefore, depending on each specific case, the appropriate approximation methods are chosen. The simulation results for the cases that cause the issues in image-based visual servoing show the strengths and drawbacks of each method, thereby helping to select the appropriate methods.