Simultaneous Localization and Capture with velocity information

Abstract

This paper introduces a method to monitor and capture the velocity, location, and dynamic behavior of a human with a self motion capture suit. To track the location of the subject in motion, a method for monitoring the velocity of a reference root point on the subject is proposed. Velocity from lower limb kinematics and the integration of accelerations are fused through Kalman Filters to achieve smooth, accurate and drift free estimation results. Using this velocity estimation and lower limb kinematics, the spatial location of the subject can be calculated. Based on some preliminary experimental results of walking motion, the RMS error within 0.48% of the velocity is achieved with respect to the commercial optical motion capture system, Motion Analysis. The localization results from the proposed method also tally well with the references. Motions like walking, jumping, running and one leg jumping are properly captured with this self-contained system. Compared with the lab-based systems, since this system does not depend on external devices and has little constraint to the human movements, it can show advantage in everyday practice and sport training in large area or home environments.