Human Motion Regeneration using Sensors and Vision

Abstract

This paper explores the opportunity of using tri-axial wireless accelerometers and optical camera system for human motion regeneration. A motion analysis system based on stick figures and musculoskeletal model is developed. Accelerometers are placed on joints of human body and markers are placed on the similar joints for the vision system. Data are acquired from these sensors in soft-real time using virtual instrumentation; the acquired data is then conditioned and converted into required parameters for motion regeneration. In vision based data acquiring is based on motion capture (MOCAP) data which is required during both the modeling and simulation process of the musculoskeletal models. While sensor based motion regeneration in this paper deals with the motion of the human body segment in the X and Y direction, looking into the motion of the anterior/posterior and lateral directions respectively, vision based motion regeneration addresses three dimensional full body musculoskeletal model. In order to regenerate specific human motion, lawn bowling action and soccer actions have been selected during the simulation.