Wireless sensor networks and video analysis for scalable people tracking

In this paper we present a system for indoor people tracking based on the combination of wearable sensors and a video analysis module. The sensor consists of an inertial platform, which provides attitude and acceleration data with a high rate. Data is fused by an Extended Kalman Filtering (EKF) to reconstruct the attitude and the accelerations experienced by the wearable sensors. The information is then integrated to reconstruct the position of the target. The presence of noise determines a gradual degradation of the localization accuracy. For this reason, a second EKF is used to reduce the uncertainty of the position by fusing the current estimation with measurements returned by the cameras.