Index of Supplementary Files from "Next2You: Robust Copresence Detection Based on Channel State Information"

This record serves as an index to the other dataset releases that are part of the paper "Next2You: Robust Copresence Detection Based on Channel State Information" by Mikhail Fomichev, Luis F. Abanto-Leon, Max Stiegler, Alejandro Molina, Jakob Link, and Matthias Hollick in ACM Transactions on Internet of Things, Volume 3, Issue 2. 2022. 

We have chosen to split the dataset into several parts to meet Zenodo size requirements and make it easier to find specific pieces of data. In total, the following datasets exist:

Raw data: This dataset contains raw channel state information (CSI) data in terms of CSI magnitude and phase values. The data is collected in the following environments: office, urban apartment, rural house, moving and parked cars, as well as additional setups in these environments such as heterogeneous devices (i.e., Nexus 6P and Raspberry Pi), different frame types to extract CSI data from (i.e., beacon), and varying transmission power of devices sending frames from which CSI is extracted. The data collection in additional setups was performed in the office environment.

These raw CSI data can be used to repeat our own experiment or to develop new context-based copresence detection schemes. To collect the raw CSI data, we used the following Android app and scripts on the Raspberry Pi 3 Model B+ (both include Nexmon patches to enable CSI extraction).

Results Data: The results dataset contains the evaluation results (e.g., computed error rates and AUCs for different cases, Right for the Right Reasons trained models, etc.). The codebase to generate these results can be found in the source code repository.