A Data Fusion Approach for Improved Positioning in GSM Networks

Reliable position estimation of a Mobile Station (MS) in a GSM network is important for many applications. Because of the diverse environments including urban, rural, outdoor and indoor and the communication signals which are usually not designed for positioning, it is difficult to obtain an accurate position estimation by using only one type of measurement. Therefore, data fusion solutions which integrate two or more types of measurement have been proposed to provide position estimation with better accuracy, reliability and coverage. In this paper, a data fusion method of integrating two different types of measurement in GSM networks, namely Timing Advance (TA) and Received Signal Strength (RSS), by using an Extended Kalman Filter (EKF) to estimate the MS's position is addressed. Several simulations based on synthetic data reveal that the estimator can track the real state under different movement situations, and that the estimated position components converge quickly to the reference values where the simulated scenario fits the model. Comparing the estimation results of the proposed method with the methods of using only TA measurements and using only RSS measurements, respectively, the data fusion method yields improved positioning accuracy.