Multiple Human Tracking with Wireless Distributed Pyroelectric Sensors

摘要

This paper presents a wireless distributed pyroelectric sensor system, whose sensing visibilities are modulated by Fresnel lens arrays, for tracking multiple humans. The aim of this study is to design and develop a wireless infrared sensor system which can track multiple humans with better data-to-object association, lower power consumption/cost/detectability, and less privacy infringement than its infrared video-based counterpart. The concept of a geometric sensor is discussed and utilized in local/global visibility modulation for the node-centric sensor system, to implement the process of data-object-association, performed by a Bayesian joint probabilistic data association scheme with validation gates. An Expectation-Maximization-Bayesian tracking scheme, consisting of detection, localization, filtering, and prediction, is proposed and implemented among slave, master, and host modules of a prototype sensor system. Experimentally the prototype system was able to simultaneously track two individuals in both follow-up and crossover scenarios with average tracking errors less than 0.5 m. The proposed wireless distributed infrared sensor system can not only run as a stand alone inmate/patient monitoring system under all illumination conditions, but also serve as a complement for conventional video and audio human tracking systems.