Our datasets are collections of channel state information (CSI) measurements created with ESPARGOS for various interesting environments. With our datasets, you can benchmark your CSI-based algorithms without the hassle of setting up experiments and collecting data. Once your algorithms are verified to work on the datasets, you can easily switch over to a real-time implementation with actual ESPARGOS hardware.
ESPARGOS datasets build on extensive experience providing datasets with DICHASUS, which you may want to consider if you need even higher-quality CSI datasets.
Larger combined antenna array, indoor lab room with metal wall, LoS and NLoS areas
Four ESPARGOS arrays are combined into one large array with 8 by 4 antennas. A metal wall blocks the LoS path in parts of the measurement area. The wall is later removed.
Four antenna arrays in indoor lab room
Four antenna arrays are pointed at a small measurement area in a lab room. Only line-of-sight channels.
Channel Charting with ESPARGOS
Indoor localization with Siamese Neural Network-based and Triplet Neural Network-Based Channel Charting. Also contains code for triangulation-based localization.
View on GitHub
This research paper describes what goes into producing one of our ESPARGOS datasets. It also briefly explains how to apply Channel Charting to one of the datasets.
Read on arXiv
Paper on ESPARGOS Hardware Architecture
This research paper explains how the hardware architecture of ESPARGOS achieves phase-coherence. Despite describing an older hardware revision, the basic explanations and results still apply.
Read on IEEE Xplore