A Fast Single-Site Fingerprint Localization Method in Massive MIMO System

Abstract

Fingerprint localization(FL) is one of the most efficient positioning scheme which exploits the characteristics of the received signal or channel information to estimate the physical position. Although there are many available positioning techniques, most of them are used in indoor positioning. In this paper, we discuss a possible method to locate a mobile device in massive multiple-in-multiple-out(MIMO) systems which represents a leading 5G technology candidate. In offline phase, the fingerprint matrix based on angle-delay channel power is extracted and compressed by three tuple(TT) method before stored into database. In online phase, coarse classification and locality sensitive hashing (LSH) are used to process the data and obtain candidate reference points (RPs). Then the weighted K nearest neighbors (WKNN) is applied to get the estimated location. The simulation results show that the proposed method has advantages of low latency and high localization accuracy compared with traditional algorithms.