Massive MIMO Channel Measurement Data Set for Localization and Communication

TL;DR

This paper introduces a comprehensive Massive MIMO channel measurement data set with location labels, enabling research in localization and communication, and demonstrates its use with CNNs for user scheduling to improve spectral efficiency.

Contribution

It provides a detailed hardware/software design of a MaMIMO testbed and releases a labeled CSI data set for research in localization and communication.

Findings

CNN-based localization achieves accurate position estimates

Data set enables joint communication and sensing research

User scheduling based on localization improves spectral efficiency

Abstract

Channel state information (CSI) needs to be estimated for reliable and efficient communication, however, location information is hidden inside and can be further exploited. This article presents a detailed description of a Massive Multi-Input Multi-Output (MaMIMO) testbed and provides a set of experimental location-labelled CSI data. In this article, we focus on the design of the hardware and software of a MaMIMO testbed for gathering multiple CSI data sets. We also show this data can be used for learning-based localization and enhanced communication research. The data set presented in this work is made fully available to the research community. We show a CSI-based joint communication and sensing processing pipeline can be evaluated and designed based on the collected data set. Specifically, the localization output obtained by a convolutional neural network (CNN) trained on the data sets is used to schedule users for improving the spectral efficiency (SE) of the communication system. Finally, we pose promising directions on further exploiting this data set and creating future data sets.