FAQ

How can sound-based sleep tracking be accurate even without wearing any device?

Sleep sounds contain rich information about sleep than biosignals that can be measured with a smartwatch. During sleep, we breathe unconsciously.

Breathing during sleep is regulated by the autonomic nervous system, and monitoring breathing sounds can measure the activation of autonomic nervous system that changes according to sleep stages. In addition, sleep sounds can indicate how much a subject tosses and turns while sleeping. Asleep’s Sleep AI integrates all of this information to enable highly accurate sleep tracking.


Does accuracy vary by gender, age, height, BMI, and disease?

Asleep's sleep AI has been trained to operate with high accuracy regardless of gender, age, height, BMI, or disease status. It has learned the characteristics of various people through 7,000 sleep polysomnography hospital data and 700,000 home environment data and continues to learn more data from more people through its service.


Does accuracy vary depending on the location where the device is installed?

The sound volume decreases as the distance increases. Therefore, if the device is too far away, the accuracy may decrease. The maximum stable distance is 2m, and it can be used up to 4m with some performance degradation. If the device microphone is positioned towards the target and there are no obstacles blocking it, it can operate stably even at greater distances.


Does it work well in a home environment with various noises?

Asleep Sleep AI works well in a typical noisy home environment. In a paper published in an international journal in April 2023, Asleep presented that Asleep’s Sleep AI can analyze sleep with high accuracy in a variety of noisy home environments. In various noisy home environments with home appliances such as air conditioners, fans, refrigerators, and other ambient noises, Asleep Sleep AI (compared to the results of home polysomnography) achieved similar accuracy to the results of sleep analysis in a hospital environment, medically verifying that Asleep Sleep AI works well in a home environment with various noises.


Reference


Are there any data security or privacy concerns?

Data security is maintained at the highest level because Asleep API is serviced by AWS, which provides the highest level of security, according to a security scheme designed by our internal security team.


In addition, sensitive data is processed on the user's smartphone and is not uploaded to the server, so it is free from privacy issues. In other words, the sound recorded by the microphone is converted into a Mel Spectrogram on the smartphone and sent to the server for analysis. Mel Spectrogram is data that has been compressed to the minimum amount of information for sleep analysis by removing sensitive information such as human voice and conversation content. Therefore, it is impossible to restore it to the original sound source. This guarantees no privacy concern, and several big companies (SKT, LG Electronics, etc.) that are very sensitive to the issue of data security and privacy are also adopting Asleep API.


Can it measure two or more people?

Currently, the multi-person measurement feature is only available as a beta service, and the accuracy level of single-person measurement may not be guaranteed in commercial services. Based on Asleep’s distance-based sound separation algorithm, we are conducting internal tests for two people using their own smartphones to measure sleep or two people using one smartphone to measure their sleep too, and the multi-person measurement feature will be applied to commercial services after medical verification.

All about Sleep, Tech, and Life

If you want to learn more about Asleep?

All about Sleep, Tech, and Life

If you want to learn more about Asleep?


Privacy Policy

Asleep Inc. 6F, 415, Teheran-ro, Gangnam-gu, Seoul, Republic of Korea


@Copyright 2023. All Right Reserved.


Privacy Policy

Asleep Inc. 6F, 415, Teheran-ro, Gangnam-gu, Seoul, Republic of Korea

@Copyright 2023. All Right Reserved.