CSI-Based Localization with CNNs Exploiting Phase Information

TL;DR

This paper explores using Channel State Information (CSI) with CNNs for localization, addressing phase inconsistencies and proposing methods to improve fingerprint reliability and CNN structure for better accuracy.

Contribution

It introduces two methods to produce reliable CSI fingerprints with phase information and analyzes CNN pooling strategies for enhanced localization performance.

Findings

Pooling over subcarriers improves accuracy more than over antennas.

Proposed methods mitigate phase inconsistencies in CSI.

CNN structure impacts localization effectiveness.

Abstract

In this paper we study the use of the Channel State Information (CSI) as fingerprint inputs of a Convolutional Neural Network (CNN) for localization. We examine whether the CSI can be used as a distinct fingerprint corresponding to a single position by considering the inconsistencies with its raw phase that cause the CSI to be unreliable. We propose two methods to produce reliable fingerprints including the phase information. Furthermore, we examine the structure of the CNN and more specifically the impact of pooling on the positioning performance, and show that pooling over the subcarriers can be more beneficial than over the antennas.