A dual-typed and omnidirectional infrared perceptual network for indoor human tracking

Abstract

This paper proposes a dual-typed and omnidirectional infrared perceptual network for indoor human target location and tracking. Two types of infrared sensors, pyroelectric infrared sensors and thermopile array sensors, are used in the sensor network for side-view and top-view perception respectively. The sensor nodes are deployed to construct an omnidirectional sensing model for detecting human targets in an indoor scenario with irregular boundary, furniture and other obstacles. The improved credit location algorithm and the adaptive threshold algorithm are applied for the human location. The location ambiguity existing in the actual situations is analysed and the compensation methods are designed by using an equivalent measurement line and a state machine to reduce the location ambiguity. Finally, a fuzzy adaptive CS-jerk (current statistical-jerk) algorithm is applied for tracking the human target. The tracking results indicate that the dual-typed and omnidirectional system works well in the indoor environment.