We design and implement a contact-free elderly-focused health monitoring system to simultaneously detect the human presence and monitor the detailed respiration status.
Human Presence Detection
Our goal is to detect whether there is a person in the room,moving or stationary. The intuition behind the Doppler spectrum is the difference of frequency spread between the human presence and absence scenario.
In this study,we obtain Doppler spectrum by computing FFT over a certain time window ,then,we extract features quantifying spectrum spread for classiﬁcation,at last,we train the Naive Bayes classiﬁer to distinguish human presence: empty room, static target, dynamic target.
The system achieves high accuracy on all three human presence cases (empty, static, dynamic), which is 99.6%, 93.6%, 98.9%.
The accuracy of detecting static cases improves signiﬁcantly with larger window size (from 88.1% with 5s window to 96.5% with 20s window).
Respiration Status Monitoring
Respiratory rate is an important indicator to measure human health, which can reflect a person’s physical state. Common respiratory rate detection mainly relies on specific sensors, which usually costs a lot and interferes with the user’s normal life.
This system only uses the existing Wi-Fi equipment and commercial wireless network card to obtain the status information of the wireless channel, and completes senseless fine-grained human respiration monitoring while establishing data communication between the two ends of the wireless receiver and receiver.
The basic idea of the system is to detect the periodic changes of the fine-grained information at both receiving and receiving ends, and complete the respiratory monitoring through signal processing.
(a)Ataxic Respiration; (b) Cheyne–Stokes respiration Based on the real-time respiratory waveform, the respiratory rate can be detected, and apNEA and other diseases can be classified and pre-diagnosed according to different abnormal respiratory waveform, which can be used for auxiliary medical treatment.
 Zhihong He, Lingchao Guo, Zhaoming Lu, Xiangming Wen, Wei Zheng, Shuang Zhou. (2019, May). Contact-free in-home health monitoring system with commodity wi-fi. In 2019 IEEE International Conference on Communications Workshops (ICC Workshops) (pp. 1-6). IEEE.
 Lingchao Guo, Zhaoming Lu, Shuang Zhou, Xiangming Wen, Zhihong He. Emergency Semantic Feature Vector Extraction from WiFi Signals for In-Home Monitoring of Elderly. in IEEE Journal of Selected Topics in Signal Processing, (under review).