Performance evaluation of MDT assisted LTE RF fingerprint framework

This paper presents a performance evaluation of the Radio Frequency (RF) fingerprinting framework for positioning using Minimization of Drive Testing (MDT) measurements specified in LTE (Long Term Evolution) Release 10. Methods based on Mahalanobis distance and Kullback-Leibler divergence are used for estimating the geographical locations. Performance evaluation of the framework is verified with system simulator by generating MDT measurement data in intra- and inter-frequency LTE network deployments with rural and urban configurations. The focus of the research is to understand how RF fingerprint positioning accuracy changes when taking into account the prospects and constraints of the RF fingerprints constructed from the MDT measurement data. The results of the study indicate that the performances of Mahalanobis distance and Kullback-Leibler divergence based methods are quite similar in rural deployments. However, in denser network deployments, Kullback-Leibler divergence based method tends to be much better for both intra- and inter-frequency environments, resulting in better than 21 meters and 38 meters positioning accuracies for 68 percent of MDT samples in studied urban scenarios, being well below the E911 emergency positioning requirements.