Accurate Radar-Based Detection of Sleep Apnea Using Overlapping Time-Interval Averaging

Radar-based respiratory measurement is a promising tool for the noncontact detection of sleep apnea. Our team has reported that apnea events can be accurately detected using the statistical characteristics of the amplitude of respiratory displacement. However, apnea and hypopnea events are often followed by irregular breathing, reducing the detection accuracy. This study proposes a new method to overcome this performance degradation by repeatedly applying the detection method to radar data sets corresponding to multiple overlapping time intervals. Averaging the detected classes over multiple time intervals gives an analog value between 0 and 1, which can be interpreted as the probability that there is an apnea event. We show that the proposed method can mitigate the effect of irregular breathing that occurs after apnea / hypopnea events, and its performance is confirmed by experimental data taken from seven patients.

Noncontact Detection of Sleep Apnea Using Radar and Expectation-Maximization Algorithm

Sleep apnea syndrome requires early diagnosis because this syndrome can lead to a variety of health problems. If sleep apnea events can be detected in a noncontact manner using radar, we can then avoid the discomfort caused by the contact-type sensors that are used in conventional polysomnography. This study proposes a novel radar-based method for accurate detection of sleep apnea events. The proposed method uses the expectation-maximization algorithm to extract the respiratory features that form normal and abnormal breathing patterns, resulting in an adaptive apnea detection capability without any requirement for empirical parameters. We conducted an experimental quantitative evaluation of the proposed method by performing polysomnography and radar measurements simultaneously in five patients with the symptoms of sleep apnea syndrome. Through these experiments, we show that the proposed method can detect the number of apnea and hypopnea events per hour with an error of 4.8 times/hour; this represents an improvement in the accuracy by 1.8 times when compared with the conventional threshold-based method and demonstrates the effectiveness of our proposed method.