Ubiquitous human motion capture using Wearable micro-sensors

Abstract

We have presented our work in progress on Human Body Motion Capture using wearable mirco-inertial sensors only. Novel drift model are proposed to deal with inherent bias problem of inertial sensors, and Extended Dynamic Bayesian Network (EDBN) are employed to model the constraints among human body segments. Given the uncertainties of human movements, multiple models are proposed to describe the behaviors of each body segment in the EDBN. The future work will be on consummating our current work. More comprehensive experiments will be further studied to evaluate our proposed method.