Datasets
Standard Dataset
Wi-Fi CSI Sensing for Sleep Disturbances in the Care of Older People
- Citation Author(s):
- Submitted by:
- Aaesha Alzaabi
- Last updated:
- Tue, 09/17/2024 - 05:36
- DOI:
- 10.21227/4rwc-1714
- Data Format:
- License:
- Categories:
- Keywords:
Abstract
This dataset contains Wi-Fi sensing data using Channel State Information (CSI) for various sleep disturbance parameters, from respiratory disturbances, to motion-based disturbances from posture shifts, leg restlessness and confusional arousals.The Wi-Fi CSI data was collected using the Wi-Fi module on the ESP32 Microcontroller units using the esp32-csi-tool.The Wi-Fi CSI respiratory disturbance data is accompanied by respiration belt data taken with the Wi-Fi measurements simultaneously using the Neulog NUL-236 respiration belt logger as ground truth. The motion-based sleep disturbances are accompanied by a timestamped actionsheet. The dataset was collected from experiments designed to investigate the monitoring sleep-disordered breathing and disturbances using Wi-Fi CSI-based non-contact sensors in the context of care of older people.
Experimental Set up
To collect Wi-Fi CSI data from the ESP32, we used the esp32-CSI-tool with two ESP32-DevKitC-VE as transceivers, one that is programmed as a transmitter (Tx) and the other as a receiver (Rx) using the built-in omnidirectional PCB antenna at 2.4 GHz. The transmitter sampling rate is set to 120 packets per second. The Wi-Fi transceivers are placed 1.6 m apart, the appropriate distance across the width of a UK double bed, in a bedroom studio environment. The unobtrusive sensor testbed is located in the Living Lab of Integrated Technologies of Care of the Advanced Care Research Centre in the Scottish Microelectronic Centre of the University of Edinburgh.
Ground Truth:
The Neulog NUL-236 respiration belt logger is worn by the subject around the waist to monitor the respiration simultaneously as a ground truth. For experiments with heart rate, Neulog NUL-208 Heart Rate and Pulse logger was used to measure heartrate. The signal sampling for both devices is set at 50 samples per second.
Additionally, for consistency measures, the subject is asked to follow the beats of a metronome set to double the intended respiration rate and is asked inhale and exhale with alternate beats for respiratory based experiments.
For motion based experiments, the subject follows a motion protocol, and a video with time stamps is taken. The time-stamped motion sheet is provided in lieu of video for privacy.
Vital Signs
The subject lies in a supine position for 5 minutes with the breathing rate instructed at 12 BPM. The duration of the experiment is 5 minutes. The corresponding CSI data, respiratory belt data and heart rate data is provided in .CSV files.
Sleep Apnoea
The subject lies in a supine position for 5 minutes with the base breathing rate instructed at 12 BPM. Five 10-second apnoea periods are instructed at 0:30, 1:30, 2:30, 3:30 and 4:30. The experiment includes central sleep apnoea (CSA) and obstructive sleep apnoea (OSA). The corresponding CSI data, respiratory belt data is provided in .CSV files.
Posture Shifts
The subject is instructed to change posture while lying down in fixed time intervals. From supine, to turing left, then supine, then right, then supine again resulting in 4 posture shifts 10 seconds apart. The experiment s repeated twice. The corresponding CSI data, and time-stamped action sheet is provided.
Leg Restlessness
The subject is instructed to raise a leg by sliding the feet up the bed toward the abdomen and remain in position for 10 seconds. The subject is then instructed to slide their leg back down and remain still for another 10 seconds before repeating the sequence for the alternate leg. This set of experiments is carried out in a supine position for a duration of 5 minutes. The corresponding CSI data, and time-stamped action sheet is provided.
Confusional Arousals
The subject gets up from a supine position and sits on the side of the bed for some time before lying down again. This is a common movement pattern in confusional arousals. This experiment is repeated for three times. The corresponding CSI data and the time-stamped action sheet is provided.
Signal Extraction
To extract the Wi-Fi CSI signal from the .CSV files, .py python notebooks for each have been provided to extract the data from the files and visualise them.