Empirical-based analysis of a cooperative location-sensing system

We have designed a novel positioning system, the Cooperative Location-sensing system (CLS) that employs the peerto-peer paradigm and a probabilistic framework to estimate the position of wireless-enabled devices in an iterative manner without the need for an extensive infrastructure or timestrenuous training. CLS can incorporate signal-strength maps of the environment to improve the position estimates. Such maps have been built using measurements that were acquired from Access Points (APs) and peers during a training phase. This paper makes three important contributions. First, it uses a particle-filters-based framework to model theoretically CLS. Second, it proposes new algorithms that incorporate real-life signal strength measurements from (APs) and peers to estimate position and distance. Third, it evaluates the performance of CLS via real-life measurements and extensive simulation, and compares it with other positioning systems. We have implemented and evaluated the CLS prototype along with its variants using IEEE802.11 and Bluetooth, and compared its performance with other positioning systems.