Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

A CSI Dataset for Wireless Human Sensing on 80 MHz Wi-Fi Channels

Francesca Meneghello, Nicolò Dal Fabbro, Domenico Garlisi, Ilenia Tinnirello, Michele Rossi

TL;DR

This work introduces a large, highly diverse CFR dataset for Wi-Fi sensing on $80$ MHz IEEE 802.11ac channels to address domain-adaptation challenges in human sensing. By spanning multiple environments, days, hardware, and including obstructed-path and semi-anechoic measurements, the dataset supports AR, PI, and PC tasks with over $13$ hours of data ($23.6$ GB) and $M=242$ sub-channels. It provides unique elements such as time diversity, hardware diversity, and obstruction scenarios, enabling robust, domain-agnostic learning and reproducible benchmarking. The dataset is publicly available on IEEE DataPort and is complemented by analysis guidance and potential future work on localization using ground-truth modalities.

Abstract

In the last years, several machine learning-based techniques have been proposed to monitor human movements from Wi-Fi channel readings. However, the development of domain-adaptive algorithms that robustly work across different environments is still an open problem, whose solution requires large datasets characterized by strong domain diversity, in terms of environments, persons and Wi-Fi hardware. To date, the few public datasets available are mostly obsolete - as obtained via Wi-Fi devices operating on 20 or 40 MHz bands - and contain little or no domain diversity, thus dramatically limiting the advancements in the design of sensing algorithms. The present contribution aims to fill this gap by providing a dataset of IEEE 802.11ac channel measurements over an 80 MHz bandwidth channel featuring notable domain diversity, through measurement campaigns that involved thirteen subjects across different environments, days, and with different hardware. Novel experimental data is provided by blocking the direct path between the transmitter and the monitor, and collecting measurements in a semi-anechoic chamber (no multi-path fading). Overall, the dataset - available on IEEE DataPort [1] - contains more than thirteen hours of channel state information readings (23.6 GB), allowing researchers to test activity/identity recognition and people counting algorithms.

A CSI Dataset for Wireless Human Sensing on 80 MHz Wi-Fi Channels

TL;DR

This work introduces a large, highly diverse CFR dataset for Wi-Fi sensing on MHz IEEE 802.11ac channels to address domain-adaptation challenges in human sensing. By spanning multiple environments, days, hardware, and including obstructed-path and semi-anechoic measurements, the dataset supports AR, PI, and PC tasks with over hours of data ( GB) and sub-channels. It provides unique elements such as time diversity, hardware diversity, and obstruction scenarios, enabling robust, domain-agnostic learning and reproducible benchmarking. The dataset is publicly available on IEEE DataPort and is complemented by analysis guidance and potential future work on localization using ground-truth modalities.

Abstract

In the last years, several machine learning-based techniques have been proposed to monitor human movements from Wi-Fi channel readings. However, the development of domain-adaptive algorithms that robustly work across different environments is still an open problem, whose solution requires large datasets characterized by strong domain diversity, in terms of environments, persons and Wi-Fi hardware. To date, the few public datasets available are mostly obsolete - as obtained via Wi-Fi devices operating on 20 or 40 MHz bands - and contain little or no domain diversity, thus dramatically limiting the advancements in the design of sensing algorithms. The present contribution aims to fill this gap by providing a dataset of IEEE 802.11ac channel measurements over an 80 MHz bandwidth channel featuring notable domain diversity, through measurement campaigns that involved thirteen subjects across different environments, days, and with different hardware. Novel experimental data is provided by blocking the direct path between the transmitter and the monitor, and collecting measurements in a semi-anechoic chamber (no multi-path fading). Overall, the dataset - available on IEEE DataPort [1] - contains more than thirteen hours of channel state information readings (23.6 GB), allowing researchers to test activity/identity recognition and people counting algorithms.
Paper Structure (11 sections, 3 figures, 2 tables)

This paper contains 11 sections, 3 figures, 2 tables.

Table of Contents

  1. Introduction
  2. Background and Existing Datasets
  3. Comparison against other public datasets
  4. Experimental Setup
  5. Wi-Fi network setup and CFR data collection
  6. Hardware specifications
  7. Measurement setup
  8. Data Description
  9. Dataset Organization
  10. Examples of Use
  11. Conclusions

Figures (3)

  • Figure 1: Semi-anechoic chamber. The Tx, Rx and monitor Asus routers used for data transmission and CFR collection are indicated in the picture.
  • Figure 2: Device and user's positions in the monitored environment for (a) bedroom with a wood bookcase in the middle; (b) living room, kitchen, laboratory, office, semi-anechoic chamber, and meeting room. Txj, Rxj and Mj, with j$\in \{1, 2,3, 4\}$, denote the transmitter, the receiver and the monitor positions, respectively. The activities were performed within the colored areas as indicated in Table \ref{['tab:configs']}. The chairs have been used for the still and sitting down/standing up activities. The other in-place activities were performed in the position of the dark-colored circles.
  • Figure 3: Example of CFR amplitude collected in the empty bedroom on two different days (AR1a_E, AR1b_E) and Pearson coefficient computed between each pair of traces collected in the empty bedroom on different days (AR sets).