The study of 3D reaching tasks on visual serving using quaternion

The image-based monocular visual servoing is difficult to realize the grasping task on three-dimension. This study presented a method of quaternion on 3D reaching tasks of visual servoing. This method used a camera and a laser scanner to acquire the information of the images and the depth. Then the parameters of Jacobian matrix could be determined by the method of the quaternion. A particular constraint mechanism was also proposed to optimize the reaching trajectory of the robot. The simulation results demonstrated our proposed method could implement the 3D reaching tasks successfully and had a significant performance improvement compared to conventional methods.