Hand Trajectory Extraction of Human Assembly Based on Multi-Leap Motions

Abstract

For the purpose of extracting the assembly motions of workers in 3C (Computer / Communication / Consumer Electronic) assembly lines and meet the needs of light weight, portability and non-contact measurement, this paper proposed a low-cost 3D motion extraction system using Leap Motions to obtain the six-dimensional (position and orientation) trajectory of the assembly motion of human hands. The multi-sensor layout is designed by analyzing the characteristics of assembly motions. To tackle the occlusion problem of Leap Motion, a multi-sensor data fusion method is presented by which the tracking precision of assembly motion in the range of target visual field is also improved. Finally, the six-dimensional trajectory extraction is verified by a simple assembly task. And the recurrence of the human assembly task is performed by UR5 (Universal Robot 5 Kg) in simulation environment of V-REP (Virtual Robot Experimentation Platform).