Performance evaluation of simplified matching algorithms for RF fingerprinting in LTE network

Abstract

Performance evaluation of fingerprinting using simplified statistical matching algorithm with two similarity metrics is presented. The collection of RF information can be automatically accomplished with the minimization of drive testing (MDT) in long term evolution (LTE) networks. In the case of non-ideal cell detection, the proposed algorithm greatly decreases the calculation of RF information and the time of position estimation. We employed the proposed algorithm in LTE cellular networks including rural, urban and Hetnet cases. The results show that the proposed algorithm reduce by 76%, 34% and 70% in computation time in rural, urban and Hetnet case respectively, while the positioning errors (PEs) have been improved in comparison with the classic KNN algorithm.