Kun Xiao

Kun Xiao

Auto-generated profile, based on publicly available information

Kun Xiao

Ningbo Konmilong Medical Apparatus Company, Ltd., Ningbo, China

Research Focus & Metrics

Academic Career Length

4Years (2023-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
19 October 2023 in IEEE Transactions on Instrumentation and Measurement
Verified
Journal Article
Closed access
KWKaikai Wang
Kaikai Wang
BFBaoliang Feng
Baoliang Feng
DZDazheng Zhao
Dazheng Zhao
KXKun Xiao
Kun Xiao
GWGang Wu
Gang Wu
CSChangcheng Shi
Changcheng Shi
Unconstrained vital sign monitoring technology has been a hot topic of research in recent years as people increasingly focus on their health status and sleep quality. Conventional contact electrocardiogram (ECG) measurement technology has difficulties in meeting the application needs of sleep scenarios due to short validity period and low comfort. To handle this problem, ECG measurement using non-contact capacitively coupled sensing technology, which do not require direct skin contact for ECG signal acquisition, could be an alternative. However, this method requires fixing the ECG measurement electrodes in two specific positions, which makes it difficult to meet the needs for body differences and multiple sleep postures. This study presents a non-contact sleep ECG measurement system based on a flexible electrode array, which can acquire ECG signals through pajamas and bed sheets using the principle of capacitive coupling. The experiment results showed that the average signal-to-noise ratio (SNR) of the data measured in the supine and lateral postures of subjects was above 30 dB, and the respiratory waves (RW) with the same frequency as contact ECG measurement system could be extracted. We also used this system to acquire ECG signals from multiple points on the dorsal of four subjects and analyzed their variation patterns, and the experimental results were in high agreement with the ECG potential employment recorded by 32-lead body surface maps. Meanwhile, the T-wave, R-wave, S-wave peaks and the ratio of T and R waves of the collected ECG signals were extracted as features. Based on this, the action classification model was trained using long short-term memory network (LSTM) to classify the postures of four subjects in supine, left lateral and right lateral positions, and the classification accuracies were 95.83%, 98.94%, 98.80% and 98.30%, respectively. Although our prototype system must be further improvement and development, these results provide a strong basis for future realization of electrode array-based ECG signal selection algorithms. The proposed system has great potential value in detecting abnormalities of ECG, heart rate (HR), respiration and other vital indicators during human sleep and realizing personalized medical treatment.