We present the rationale, architecture, and implementation of an intelligent room, built upon a dense radio-based network of diverse electronic sensors, and interfaced with a cluster of dedicated computers. This room (an "iRoom") is capable of high-order tasks in event detection, analysis, processing, deduction, and inference. Initial focus has been to provide answers to the "Five W's" regarding a human visitor: Who, What, When, Where, and Why. The foundation technique that underpins this research is Sensor Fusion: the mathematical combination of disparate sensor reports and analyses into a rational result which is more accurate and more useful than its constituent parts.

We describe a method of using a network of thermopile sensors distributed along the walls of a room to locate a person within the room. At any given time, a person would be in the view of at least one of the sensor nodes. An algorithm is used to calculate the distance of a person from each sensor node based on its temperature reading and locate the person based on a combination of the sensor readings from the distributed nodes. This method is immune to lighting changes and works even in total darkness. Furthermore, this method of sensing can also detect events such as a person walking past the doorway to a room, lingering outside, and entering or leaving the room.