Research in the International Journal of Sensor Networks describes a new way to monitor human sleep that relies entirely on the Wi-Fi signals in the home. It promises an entirely non-intrusive and yet accurate alternative to conventional techniques.
A good night’s sleep, night after night is, for most of us, is a fundamental part of good health, optimal mental performance, and emotional well-being. Understanding sleep and its insidious counterpart, insomnia, however, has often relied on cumbersome devices and complex approaches to research involving polysomnography, the recording of brain waves, eye movements, muscle activity, and heart rate to assess what stage of sleep a person is in.
Polysomnography is a valid approach to sleep studies but usually requires participants in sleep experiments to spend the night in specialized facilities under clinical supervision, making it costly and impractical for long-term or widespread monitoring. Moreover, the artificial conditions are likely to disrupt the normal sleep patterns the participants would experience if they were tucked up in their own beds at home.
Wearable devices such as smartwatches and smartphone applications offer a more convenient option for sleep science. However, such devices generally track only basic metrics, such as sleep duration, and many people enlisted into studies find such devices uncomfortable to wear when they go to bed.
The new research sidesteps these limitations by exploiting Channel State Information (CSI), this is a metric derived from Wi-Fi signals that captures how electromagnetic waves propagate through space. When a person moves or changes posture during sleep, these subtle shifts alter the amplitude of the Wi-Fi signals that permeate the bedroom. By statistically modelling these variations, the researchers explain that it is possible to detect static sleep positions and nocturnal movements without the need for the subject in the study to wear specialist sensors or even their own smartwatch.
Data collection simply involves the use of a system of Wi-Fi antennae that work on a range of frequency channels. The collected signal data is then subject to a statistical analysis, a probability density function, that examines how likely different signal amplitudes are to occur when associated with a person’s movements during sleep. The team says their approach achieves recognition rates of more than 95% for common sleep positions and nocturnal movements.
The technology will be useful to sleep researchers but could be extended to monitoring older people and the mobility-impaired. The same Wi-Fi analysis might be operated in real-time to spot irregular movements, falls, or sleep disturbances and so allow for more timely interventions or assistance.
Liu, Y., Cao, Z. and Hu, M. (2025) ‘Sleep behaviour monitoring based on the probability density model’, Int. J. Sensor Networks, Vol. 48, No. 4, pp.227–240.
No comments:
Post a Comment