Probabilistic Tracking of Pedestrian Movements via In-Floor Force Sensing

Abstract

This article presents a probabilistic approach to the tracking and estimation of the lower body posture of users moving on foot over an instrumented floor surface. The latter consists of an array of low-cost force platforms providing intermittent foot-floor contact data with limited spatial resolution. We use this data to track body posture in 3D space using Bayesian filters with a switching state-space model. Potential applications of this work to person tracking and human-computer interaction are described.