Improving Sleep Quality Using Sound Intervention

Objectives
This project studies the use of interactive artificial intelligence (AI) based sound interventions to improve sleep quality in noisy home environments. Masking noise by emitting sound is well known from so called “sound noise machines”, yet current state-of-the-art machines operate by being constantly on or through a set timing function. Long term usage of these can lead to auditory perceptual problems and should only be active during the onset of awakenings and abrupt noises. Yet, detecting the onset of awakenings is currently challenging without access to highly specialized equipment and expertise for data analysis. This project explores how wearable physiological sensors and microphone arrays can be applied to detect periods of awakenings, i.e. sleep arousals, which typically denote points in our sleep where we are volatile against noise. Using AI to intervene in real-time, we can proactively apply noise masking at appropriate periods during sleep, which in turn may increase sleep quality significantly. Our research question is: How can interactive AI-based sound interventions improve sleep quality in noisy home environments?

The project can lead to a significant innovation for the partnering company SoundFocus ApS, enabling an interactive noise masking solution for people living in noisy environments. To facilitate proactive interaction towards sleep arousals, we will develop an AI-based software solution and study its effectiveness in terms of objectively detecting arousals and external noise through various physiological sensors and microphone arrays.

Success criteria
1. Using AI to detect early signs of sleep arousals through physiological sensors and micro-phone arrays in noisy home environments
2. Increasing sleep quality in noisy home environments through interactive sound interven-tion