A dataset for Wi-Fi-based human activity recognition in line-of-sight and non-line-of-sight indoor environments

Published: 16 February 2021| Version 2 | DOI: 10.17632/v38wjmz6f6.2
Contributors:
Baha' A. Alsaify,
,
,

Description

The dataset of Wi-Fi signals was captured in three indoor environments. 30 subjects participated in the data collection process. In each of the environments, the subjects were instructed to perform a set of pre-explained experiments between two devices, each of them has Intel 5300 network interface card installed. On the transmitting device one transmitting antenna was used, while on the receiver, the transmitted signals were captured by three antennas. The designed experiments were described comprehensively to each of the participating subjects before the start of the experiments. For the first and second environments, a LOS architecture was used while a NLOS architecture was used in the third environment. The Intel 5300 network interface card was configured to operate within the 2.4 GHz frequency band, wireless channel number 3, channel bandwidth of 20 MHz, and 320 packets/second sampling rate. The data collection process took place in three environments. For each of the environments, 10 subjects performed 5 different experiments. Each subject performed 20 trials for each of the experiments. The exchanged Wi-Fi signals were recorded and the RSSI value and the CSI values were recorded.

Files

Institutions

German Jordanian University, Jordan University of Science and Technology

Categories

Computer Science Applications, Signal Processing, Pattern Recognition, Human-Computer Interaction

Licence