A Calibration-Free RSS-Based Mobile Positioning System

Received power in cellular networks is commonly employed in mobile localization systems. However, uncertainties in power-distance mapping and dynamics of propagation models challenge the performance of the positioning system. Although collecting realistic data in the target area may reduce the uncertainties, it requires a time-consuming site survey and high-cost labor efforts. This study proposes a novel algorithm to enhance the performance of mobile localization without the need of additional calibration effort. The proposed algorithm utilizes the pairwise information between base stations (BSs), which is assumed to be available, and then localizes the user based on multidimensional scaling. Unlike traditional methods, the proposed approach further considers the geometric structure between BSs to compensate for the problem of distance estimation, thus achieving more accurate location estimations. We applied this approach in a realistic GSM network and experimental results demonstrate the effectiveness of our approach. The proposed algorithm outperforms previous calibration-free positioning methods, including Cell-ID and enhanced Cell-ID, in reducing the mean error by 16.74%-38.56% and 18.22%-20.96%, respectively.