Considerable works have been performed for such tasks as inserting a cylindrical peg into a hole with same type. For this task, the passive RCC device developed in MIT does work well. As an alternative the active assembly methods have been developed elsewhere.
In this thesis assembly of a block which have a rectangular cross section into a rectangular type hole is conducted. In this task, on the contrary to a cylindrical peg-in-hole, the coincidence of the shape about the rotational axis is an important factor to successful insertion.
The algorithm were proposed in this thesis under some assumptions, which position and orientate the rectangular block to coincide with the hole with a single camera. The camera and vision system should be calibrated beforehand. However, this process is not critical to the accuracy of the position and orientation in the proposed servoing algorithm. The orientation of the rectangular peg is estimated based upon the vertices of the block extracted from the processed image. The world coordinates of the vertices were found through a series of finite movement of the block.
The algorithms were applied to an experimental system with an industrial robot. Through a series of experiments it was found that guidance of the robot based upon the proposed visual servoing algorithm satisfactory result.
