Human Motion Capture Error Correction Method without Force-Feedback

PubDate: July 2019

Teams: College of Intelligent Science and Technology, National University of Defense Technology

Writers: Zhuqi Qi；Zhexue Ge；Xu Luo

PDF: Human Motion Capture Error Correction Method without Force-Feedback

Abstract

Aiming at the influence of the lack of force-feedback on the accuracy of human motion in virtual reality applications, based on the human dynamics model and Newton-Euler calculation method, the upper limb movements under different external forces were calculated, and the dynamic calculation was verified by a series of experiments. On this basis, the data fitting method was used to correct the human motion capture error caused by the deletion of force-feedback, and the third-order polynomial fitting results of human joint angle and force magnitude were obtained. Further experiments have showed that the human action correction method can effectively reduce the traditional action errors mentioned and has good application value.