With the province’s recent decision to use a contact tracing app to help curb the spread of COVID-19, a University of Guelph-led project could help in fine-tuning or improving the accuracy of this technology.
U of G engineers have developed a new smartphone app that provides more secure and accurate contact tracing vital to limiting the spread of COVID-19.
Developed using wireless communication and machine learning, the Smart Contact Tracing app ensures greater accuracy and privacy than other systems while identifying any recent contacts infected with the novel coronavirus and reminding users to maintain physical distancing, said Prof. Petros Spachos.
“The application we developed could be very useful as an upgrade to any contact tracing application available, including the one recently approved for use by the provincial government,” said Spachos.
Using machine learning, the team has made the app more accurate than other systems, he said. Apps that perform well on handheld phones or that are “visible” to each other lose accuracy when the devices are in a pocket, purse or backpack.
Using those “hidden phone” scenarios, tests of the Smart Contact Tracing app showed that accuracy improved from about 56 per cent to 87 per cent. The smart system also “learns” to distinguish when the user is at home or in another private space and stops recording contact information there.
The automated system reveals no identifying information about individuals or where interactions occurred – a key feature for many users with privacy concerns, said Spachos.
“No location data or other personal data about the user are kept. Preserving privacy is our first priority. We don’t know where the interaction took place and we don’t know who the users are.”
The app alerts users when required physical distancing is not being maintained, or whether anyone nearby is infected.
“If you’re within two metres, the phone will start vibrating.”
- Prof. Petros Spachos
A paper on the system has been accepted for publication this summer in the IEEE Systems Journal.
Automated contact tracing for COVID-19 infection using a phone app is faster and less labour-intensive than manual means of tracking down everyone who may have interacted with an infected individual, said Spachos, who worked on the project with lead author Pai Chet Ng, a student visiting U of G from Hong Kong University of Science and Technology; U of G engineering professor Stefano Gregori; and University of Toronto professor Konstantinos Plataniotis.
Various smartphone apps allowing automatic tracing have been developed worldwide during the COVID-19 pandemic. Unlike the U of G system, those systems may not ensure privacy or may lack security against hacking, said Spachos.
The Smart Contact Tracing app records a user’s contacts for the previous two weeks, including passersby with only fleeting contact.
If a user tests positive for the coronavirus, they can use the phone app to upload that encrypted information to the cloud. From there, a message alerts everyone who had contact with that individual without providing names, phone numbers or locations.
That enables recipients to pursue testing to see whether they are also infected, bypassing any need for health care professionals to undertake manual and time-consuming contact tracing.
Referring to the team’s system, Spachos said, “It’s similar to a lottery. Whoever has the numbers, it means they were in close proximity to someone who was infected. Then it’s up to the users to go get tested. We do not know who has the numbers, nor can we identify them.”
The system relies upon initial users sharing positive test results and other users pursuing testing. He said he expects users will voluntarily cooperate, particularly with privacy assured.
“COVID-related apps are extremely popular.”
The team received a $50,000 Alliance COVID-19 grant from the Natural Sciences and Engineering Research Council and $20,000 from U of G’s COVID-19 Research Development and Catalyst Fund. The researchers will work with OMESH Networks Inc., a wireless communication firm based in Toronto.
Contact:
Prof. Petros Spachos
petros@uoguelph.ca