Hybrid GPS/GSM localization of mobile terminals using the extended Kalman filter

In dense urban and indoor scenarios, the Global Positioning System (GPS) often cannot provide reliable mobile terminal (MT) location estimates, due to the attenuation or complete shadowing of the satellite signals. Cellular radio network-based localization methods, however, provide MT location estimates in almost every scenario, but they do not reach the accuracy of MT location estimates provided by GPS. Thus, a promising approach is to combine measured values from the cellular radio network and GPS, which is known as hybrid localization. In this paper, an extended Kalman filter (EKF)-based hybrid localization method is proposed that combines round trip time and received signal strength measured values available from the Global System for Mobile Communication (GSM) and pseudorange (PR) measured values from GPS, in order to track the MT's movement. In contrast to existing hybrid approaches, an EKF-based MT tracking algorithm is proposed that additionally takes into account sectorized BS antennas which are typically employed in existing GSM networks, and it takes into account PR instead of geometric range (or time of arrival) measured values from GPS as GSM is normally not time-synchronized to GPS. Simulation and experimental results show that, compared to an EKF that is based only on GSM measured values, the proposed EKF that additionally incorporates PR measured values yields improved MT location estimates.