In order to avoid the common problems of time-consuming kinematics inversion-computing and high computational complexity of vision system calibration in the autonomous gripping of robotic arm

a robotic gripping approach based on combined model of Gaussian process regression（GPR） and kernel ridge regression（KRR）models is proposed.In the learning phase

the target detection and instance segmentation algorithm should be trained based on Mask-RCNN

and the manipulator grasping strategy should be trained based on GPR and KRR.In the grasping phase firstly

target detection and instance segmentation algorithm is used to obtain the pose of the target object

then according to the mapping relationship between the pose of the target object and the joint angle of the robot arm

joint angle of the robot arm is predicted and control the robot arm to complete the grasping task.Experimental result show that the proposed method does not require the calibration of the visual system and the kinematics inversion-computing of the robotic arm

and can accurately obtain the pose of the target object.It is verified by the AUBO i5 robotic arm that this method can achieve better accurate grasp of the target object.