5G for Future Wireless Networks. Second EAI International Conference, 5GWN 2019, Changsha, China, February 23-24, 2019, Proceedings

Research Article

Smartphone-Based Intelligent Sleep Monitoring

Download
100 downloads
  • @INPROCEEDINGS{10.1007/978-3-030-17513-9_4,
        author={Pansheng Fang and Zhaolong Ning and Xiping Hu},
        title={Smartphone-Based Intelligent Sleep Monitoring},
        proceedings={5G for Future Wireless Networks. Second EAI International Conference, 5GWN 2019, Changsha, China, February 23-24, 2019, Proceedings},
        proceedings_a={5GWN},
        year={2019},
        month={4},
        keywords={Sleep scoring Sleep event detection Microphone},
        doi={10.1007/978-3-030-17513-9_4}
    }
    
  • Pansheng Fang
    Zhaolong Ning
    Xiping Hu
    Year: 2019
    Smartphone-Based Intelligent Sleep Monitoring
    5GWN
    Springer
    DOI: 10.1007/978-3-030-17513-9_4
Pansheng Fang,*, Zhaolong Ning1,*, Xiping Hu2,*
  • 1: Dalian University of Technology
  • 2: Shenzhen Institutes of Advanced Technology, Chinese Academy of Sciences
*Contact email: ps.fang@siat.ac.cn, zhaolongning@dlut.edu.cn, xp.hu@siat.ac.cn

Abstract

The sleeping quality is one of the most important factors to judge people’s health status, and has drawn increasing attention of the public recently. However, the quantified results of sleeping quality can generally be achieved in labs with the help of high precision instrument, such as Actigraphy or professional graph like Polysomnography (PSG), and are thus not available for the general public. In this paper, we construct a novel way of sleep-scoring system implanted in the iSmile app. iSmile first collects the sounds recorded by smart phone recorder, then classifies the sound frames with a light weight decision tree algorithm. Based on the number and the average amplitude of sleep-related events, we score the users’ sleeping quality in three aspects (respectively cough-score, snore-score and talk-score) using Pittsburgh Sleep Quality Index (PSQI) and Pediatric Sleep Questionnaire (PSQ). During users’ sleeping period, iSmile also collects data from the accelerator sensor to predict the users’ mood (presented in valence and arousal) and recommend smart alarm sounds to help improve their mood. For the experiment, we involved 5 participants (20 nights in total) and achieved high precision of predicting sleep events (above 89%), with the users’ valence and arousal improved by 14.57%. From succinct chart of sleeping score on the App UI, users can see the visualized results of their sleeping quality.