Indoor human localization using PIR sensors and accessibility map

Tracking human's location in indoor environments is very import, for it enables a wide range of applications related to robot service, appliance control and other smart home applications. This paper proposed a new, integrated method for indoor human localization based on passive infrared (PIR) sensors and an accessibility map of the indoor environment. The accessibility map reflects human's visiting habits in that environment, which also contains geometric as well as furniture layout information. First we build the accessibility map based on the human's visiting habits as the prior knowledge. Then the PIR sensor data is used to estimate the rough position. A more practical binary sensing model is used to model the PIR output. A particle filter fuses the PIR sensor data and the accessibility map information to refine the localization. To evaluate our method's performance, simulation experiments are carried out. The results showed that the proposed accessibility map-based PIR localization method outperforms the PIR only localization.