Kinematics Modeling and Trajectory Planning for NAO Robot Object Grasping based on Image Analysis

This paper presents a method of kinematics modeling and trajectory planning for NAO robot object grasping. In order to judge and grasp objects autonomously, this method combines the image analysis and kinematics knowledge. Firstly, the collected images by the NAO robot are processed through threshold segmentation and median filtering. According to the shape information of the object, the NAO robot determine automatically whether to adopt single-arm or double-arm grasping. Secondly, choosing the Denavit-Hartenberg (D-H) method, the kinematics model of the NAO robot's arm is built. By solving the forward and inverse solutions of kinematics, the joints variable angles of the NAO robot's arm are obtained. Then, the pose and position of each end-effector are calculated. Finally, the motion trajectory of the end-effector is planned by designing the trajectory planning in joint space. The experimental results show the effectiveness and intelligence of the method, which demonstrates that the proposed method has practical value for the research of NAO robot.