Sensor networks hold the promise of intelligent buildings, where the building adapts to the behavior of their occupants to improve the safety and the security as well productivity and efficiency.
The purpose of the Smart Home and Smart Offices is to make the inhabitants feel more convenient and more comfortable.
In this thesis we propose a human localization system based on a low power, low cost sensor network.
We describe a system that performs the localization of humans in a room.
An occupancy map provides a top view of the scene and indicates the locations occupied by people.
Such map, is of practical use in many application such as surveillance, smart rooms and care.
In order to provide this information we designed a simple sensor module based on a PIC microcontroller equipped with a thermopile array.
We propose a novel approch to select and fuse all information of interest provided by the sensors based on the Dempster Shafer Theory of Evidence.
The effectiveness of our solution has been proved by test in an office room of 5x4 meters.
We consider this work to be a first step to gain access about occupants behaviour in the context of an entire building in a way that is only minimally intrusive to the occupants privacy.