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

Abstract EN

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.