A fusion architecture for tracking a group of people using a distributed sensor network

This paper presents an architecture to fuse the information from different spatially distributed unattended ground sensors (UGSs) consisting of multiple modalities to track a group of people walking along trails in open areas. The UGS system we used consists of ultrasonic, seismic and passive infrared (PIR) sensors. Estimating the size of the group is done by counting the number of targets using ultrasonic sensors. Determining the composition of the group is done using several classifiers. The fusion is done at the UGS level to fuse information from all the modalities to determine the presence of a target(s) and the information from multiple UGS systems are fused to track a group. The algorithms are tested on data collected in the wilderness with three UGS systems separated 75 m apart on a trail.