Research On Intelligent Identification And Tracking Systems For COVID-19 Close Contacts

As the prevention and control of COVID-19(Corona Virus Disease 2019)has become normalized,the resumption of work,production and schools across the country has been accelerated.The main challenge in this process is to avoid secondary clusters of outbreaks triggered by COVID-19.Locking for close contacts has become the key to prevention and control of COVID-19.The existing methods of searching for close contacts have some shortcomings,such as the over-reliance on patient memory,the low efficiency and easy to make mistakes and omissions.In this paper,an intelligent Virtracker(Virus Tracker)system is proposed to realize one-stop intelligent tracking from the client to the management.The contact information of users is collected through the Application software of mobile phones.The contact distances among users is inferred on the cloud,and the contact network time sequence diagram is constructed to carry out the iterative transmission of infection risk to predict potential infected persons for prevention and control of COVID-19.The user contact information can be acquired by the Bluetooth Low Energy(BLE)near field communication technology in mobile phones and the key technology of distance estimation based on RSSI(Received Signal Strength Indication)is discussed in this paper.In order to improve the accuracy of distance estimation,the influence of equipment parameters and external environment on RSSI measurement is analyzed and the key problems estimating distances among large-scale mobile devices with randomly distribution are studied.In view of the large-scale variation of RSSI measurement caused by channel fading,shadow,environmental change and other factors,a deep learning distance estimation model including RSSI as the main element and other sensors as the auxiliary is proposed in this paper.A variety of sensor data are integrated to realize the state modeling of the mobile phone and reduce the influence of the state of the receiving devices on the RSSI measurement.In order to further improve the accuracy of distance Estimation,the residual idea is introduced to build a COVID-Proximity Estimation Network(COVID-Proximity Estimation Network).Based on real contact data,the simulation results show that the proposed Covid Proxi Net model significantly reduces the influence of other factors on RSSI measurements and the accuracy of distance estimation was as high as 93.86%.Finally,the characteristics of chain transmission and cross-infection of the virus are considered in this paper.The contact information is used to generate a dynamic time sequence diagram network,and the contact intensity is investigated by contact distance and contact duration.A CREM(Covid Risk Estimate Model)is built to calculate the infection risk of all users.The simulation results show that the predicted coverage rate of direct contacts is 100%,the nucleic acid detection efficiency is 40% which is higher than that of the whole detection.And the risk level of infection among intergenerational contacts is given,which is helpful for differentiated epidemic prevention and avoiding the run on medical resources.At the same time,the control-experiment shows that CREM model can effectively reduce the dependence of the system on the accuracy of distance estimation.