DeepWiPHY: Synthetic and real-world IEEE 802.11ax OFDM symbol dataset

Citation Author(s):
Yi
Zhang
The University of Texas at Austin
Akash
Doshi
The University of Texas at Austin
Rob
Liston
Cisco Systems Inc.
Wai-tian
Tan
Cisco Systems Inc.
Xiaoqing
Zhu
Cisco Systems Inc.
Jeffrey
Andrews
The University of Texas at Austin
Robert
Heath
The University of Texas at Austin
Submitted by:
Yi Zhang
Last updated:
Sat, 04/11/2020 - 01:38
DOI:
10.21227/7h3f-xc81
Data Format:
License:
0
0 ratings - Please login to submit your rating.

Abstract 

This dataset is associated with the paper entitled "DeepWiPHY: Deep Learning-Based IEEE 802.11ax Receiver". It has synthetic and real-word IEEE 802.11ax OFDM symbols. The synthetic dataset has around 110 million OFDM symbols and the real-world dataset has more than 14 million OFDM symbols. Our comprehensive synthetic dataset has specifically considered typical indoor channel models and RF impairments. The real-world dataset was collected under a wide range of signal-to-noise ratio (SNR) levels and at various locations in a large office. To create this real-world dataset, we built a passive sniffing-based data collection testbed with a commercial AP, a Samsung mobile phone and two USRPs. 

Instructions: 

The documentation and instructions will be found in the uploaded files.

The models and weights of the trained NNs will be uploaded soon.