Evaluating the accuracy of a mobile Kinect-based gait-monitoring system for fall prediction

Abstract

Accurately and pervasively monitoring the human walking pattern (or gait) is fundamental to predict falls and functional decline, which are among the leading causes of injury and death in older adults. Existing gait-monitoring devices are not routinely used in clinical practice since they lack accuracy, ease-of-use, and unobtrusiveness. We present a novel breakthrough Kinect-based robotic system to accurately monitor the human gait during normal daily-life activities. Our system combines many interesting features: it has unlimited capturing volume, it is low cost, and does not require fiducial markers on the person. We present an extensive study of its accuracy in computing fall-prediction parameters when compared to the Vicon motion-capture system.