Vision Based Object Grasping of Robotic Manipulator

Abstract

In order to better interact with the complex and unstructured environment and complete the recognition and capture of the target object, a humanoid robot arm control method based on binocular vision applied to the service robot platform is proposed. Firstly, the kinematics model is established for the manipulator, and the forward kinematics and inverse kinematics are analyzed. Target recognition method based color segmentation, then the identified three-dimensional coordinate information of the target controls the robot arm to complete the grab task. In this paper, a blind grasp is performed by the ROBOROT ARM1 manipulator. The experimental results show that the robot can complete the recognition and automatic capture of the target object.