Abstract:
This work describes a new approach for localizing people by cooperative sensor fusion of lightweight camera and wearable accelerometer measurements. We present the algori...Show MoreMetadata
Abstract:
This work describes a new approach for localizing people by cooperative sensor fusion of lightweight camera and wearable accelerometer measurements. We present the algorithm to identify people moving around as they are detected by cameras deployed in the infrastructure. The algorithm uses a correlation metric to develop an ID matching algorithm that can associate people in the scene to their global ID emitted from a wireless accelerometer sensor node worn on their belts. First we conduct a set of preliminary experiments to verify that the quantities of interest easily measurable by off-the-shelf components. We validate our metric and the performance of the proposed ID matching algorithm using simulations on testbed data that also includes a crowded scenario.
Published in: 2010 Proceedings IEEE INFOCOM
Date of Conference: 14-19 March 2010
Date Added to IEEE Xplore: 06 May 2010
ISBN Information: