Kinematic Uncertainties in Human Motion Tracking and Interaction

Abstract

Human kinematic model is crucial for precise human motion tracking and representation. While using wearable sensors to track the human motion, the human kinematics cannot be measured exactly due to the complexity of the human body and the sensor measurement errors. For the existence of these uncertain parameters in the human kinematics, the contact interactions with the environment during the activity cannot be correctly captured. Research works on the kinematic uncertainty problem for motion tracking is still lacking. In this paper, after discussing on the kinematic uncertainties, a method to re-adjust the body posture for correct motion and interaction tracking is introduced. In the preliminary experiment validation, an IMU based motion and interaction tracking system is used to capture the human motion with contact interactions in the environments. After considering the kinematic uncertainties in the tracking, motions like climbing ladders, rock climbing, gym etc., can be captured and represented with correct interactions with the environment.