Kinect-based Data Processing Noncontact Robotic Arm Control System

Abstract

We propose a method to control the motion of a robotic arm in real-time by using Kinect sensors to identify, track and simulate the motion of a human arm. This method uses Kinect’s 3D skeleton tracking technology to obtain joint data and controls the motion of the robot arm after processing the data. The biggest attempt of this research is to apply the spatial 3D shear transformation to the actual manipulator control, hoping to correct the arm data of different heights, arm lengths, and arm swing amplitudes into a unified operation unit space, and to limit the position of robot arm through the joint angle. In addition, by adding two-hand gesture recognition conditions to assist control, increase control functions as much as possible, and improve the control system. In this way, the cost of relying on external equipment in the traditional manipulator control scheme is reduced, which also proves the feasibility of such a method.