Statistical performance of subspace matching mobile localization using experimental data

In this paper, we present some statistical results on localization performance of Subspace Matching Localization (SML), a technique for localizing non-cooperative transmitters in a microcellular environment [1, 2] using one(multiple) base-station(s) that is(are) equipped with an(multiple) antenna array(s). We use extensive experimental measurements obtained in Yokosuka, Japan. SML exploits the dissimilarities in spatial multipath propagation characteristics of distinct locations, as reflected in the received respective signal subspaces to localize a transmitter. The location detector utilizes a match of the received signal subspace with those in a database. The database can be constructed by measurement campaigns, or ray tracing based simulations. In this paper, we report the localization accuracy of SML using a database that is constructed using a measurement campaign and up to two base stations. Issues regarding the variations of signal subspace due to non-stationarities in the environment and tracking of mobile transmitters are also addressed briefly.