Localization with Ambient Sensors

This paper deals with the design, implementation and evaluation of an indoor tracking system that relies solely on an infrastructure of wirelessly-connected sensors and on architectural, behavioral and contextual knowledge to locate and track a person. Such a system can be very useful when users do not want or cannot wear a transmitting device. This situation arises when it would be obtrusive to ask to wear a tag, e.g. in a home-automated apartment inhabited by people of all age that do not require special assistance, or in an elderly-assistive situation when the person is changing clothes, going to bed, taking a bath, etc. The paper fully describes the techniques used and shows some performance data collected in a laboratory environment with real sensors.