Road traffic estimation with signal matching in mobile phone using large-size database

This paper presents a road traffic estimation method using a GSM mobile phone as a probe. The technique relies on localization by signal strength matching. Our emphasis is on the unusual-large size of the database. We built a module that automatically collects location-signal with one second refresh rate and carried it along a test route to collect database. To handle big database, we additionally present a noise reduction algorithm, which eliminates bad data and compacts the database. For the localization algorithm, the fast nearest neighbor strategy is applied in the database matching to reduce computation time. The position estimations which formerly implemented noise reduction are compared with the ordinary results. Furthermore, we compared the computation time of the fast nearest neighbor algorithm to the normal method. The results show that the mobile phone positioning based on signal strength pattern matching can locate its location down to 63 meter accuracy.