Unconstrained Apnea and Asthma Symptom Detection Using Ultra Wide Band Radar

Abstract

Abstract

Unconstrained Apnea and Asthma Symptom Detection Using Ultra Wide Band Radar

MyungJun Koh

Interdisciplinary Program of Bioengineering The Graduate School Seoul National University

Respiration and coughing are evident vital signs to check ones condition are sound. Long-term respiration monitoring is needed to find out if the person is in good health condition. Long-term respiration monitoring can be cumbersome when devices are constraining subjects body. Measuring respirational signal with the unconstrained method is good for patients or babies or elders who have a lower level of enduring cumbersomeness. The purpose of the study is to prove the feasibility of Ultra Wide Band (UWB) radar for detecting abnormal respiration. UWB Radar uses 3.1 to 10.6 GHz frequency range and transmit pulse with very short duration. Impulse radiation is 50 degrees from Vivaldi antenna. Its small size and low power consumption make UWB radar application for home or inside a vehicle. Apnea and coughing data was classified from normal breathing data with classifier such as Support Vector Machine (SVM), Decision Tree (DT) and Random Forest (RF) with statistic and frequency domain features. For apnea detection, reference device data showed the best result with RF classifier with 100% mean sensitivity, 70% mean precision, 0.82 F measure and UWB radar data showed the best result with RF classifier with 100% mean sensitivity, 70% mean precision, 0.82 F measure. For coughing detection, reference device data showed the best result with RF classifier with 100% mean sensitivity, 71% mean precision, 0.83 F measure while UWB radar data showed the best result with RF classifier with 100% mean sensitivity, 83% mean precision, 0.91 F-measure. Detecting apnea and coughing while a person is moving or in different posture should be studied in future studies. This study expects more application of UWB radar in long-term vital sign monitoring.

Keyword: Ultra wide band radar, Respiration, Apnea, Cough, Asthma, Unconstrained Student Number: 2015-21207