. . . "Worcester Polytechnic Institute" . . "2021" . "2021-12-01T18:36:03.089063009-05:00"^^ . "http://rightsstatements.org/vocab/InC/1.0/" . . "Pahlavan, Kaveh" . "Agu, Emmanuel O." . "MS" . "Agu, Emmanuel O." . "Pahlavan, Kaveh" . "Etd" . "2021-11-30T12:14:32.586832427-05:00"^^ . "Electrical & Computer Engineering" . "etd-31591" . "depositor@wpi.edu" . . "2021-09-04" . "Semenov, Oleksandr" . "2021-08-23" . . "PACT" . "COVID" . "proximity detection" . "nDCF" . "machine learning" . "RSSI" . "Machine Learning Estimation of COVID-19 Social Distance Using Smartphone Sensor Data" . "Thesis" . . "COVID-19 is spread from an infected to a healthy person when they are within 6\r\nfeet from each other for longer than 15 minutes. To limit disease transmission, there\r\nis a need for technology that could identify whether subjects were near each other\r\nlonger than 15 minutes. In this thesis we systematically investigate Machine\r\nLearning (ML) methods to detect proximity by analyzing data gathered from\r\nsmartphones’ built-in Bluetooth, accelerometer, and gyroscope sensors. We show\r\nthat the proximity classification (< 6ft or not) can achieve 72%-90% accuracy using\r\nthe accelerometer, 78%-84% accuracy using gyroscope sensor, and 76%-92%\r\naccuracy with the Bluetooth radio, while sensor fusion shows accuracy as high as\r\n97%. Our model outperforms current state-of-the-art methods using neural networks\r\nand achieved Normalized Decision Cost Function (nDCF) score of 0.34 with\r\nBluetooth radio and 0.36 with sensor fusion." . . "ActiveFedora::Aggregation::ListSource" . . . . . .