Epidemic Contact Tracing With Campus WiFi Network and Smartphone-Based Pedestrian Dead Reckoning

Abstract

Due to the rapid spreading of infectious disease COVID-19, numerous campus students are increasingly exposed to the dilemma and thus the provisioning of a safe environment becomes of vital importance. As a well-established approach, contact tracing could contain epidemic diseases spread. Since WiFi network could cover almost the whole campus and each student carries at least one WiFi capable device (i.e., smartphone), in this work, an epidemic contact tracing with campus WiFi network and smartphone-based pedestrian dead reckoning (PDR) is proposed, involving not only coarse-grained duration, but also fine-grained distance between students. First, students' location distribution and duration are captured by non-perception WiFi network logs with highly flexibility. Then, the convolutional neural network (CNN) model is utilized to real-time recognize landmarks in PDR positioning trajectory, followed by the particle filter algorithm to fuse both the PDR positioning results and detected landmarks, thereby calibrating PDR cumulative error and calculating the social distance between students. Next, we analyze the contact degree between students by integrating duration and social distance. Finally, in a campus environment with an coverage of about 600m <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup> , we simulate a COVID-19 case study to validate proposed approach, showing that the average positioning error is reduced from 3.23m to 2.77m.