CMU Professor Creates Innovative App to Anonymously Trace Exposure to COVID-19
By Emily PayneMedia Inquiries
- Mellon College of Science
As the world continues to focus on slowing the spread of COVID-19 and determining how to safely restart the world economy, a team of designers, mathematicians, computer scientists and physicists led by Carnegie Mellon University Professor Po-Shen Loh, has developed an innovative new contact tracing app that uses ultrasound technology to more accurately follow people's exposure to COVID-19 than other apps that use Bluetooth alone.
The free Android app, NOVID, is currently in public beta testing and Loh is awaiting its approval in the Google Play Store, and an iOS version is under development.
"I had an idea of how to help the world return to normalcy," said Loh, a member of the Department of Mathematical Sciences faculty in CMU's Mellon College of Science. "And the Carnegie Mellon and broader Pittsburgh community had everything I needed to start making that idea into reality."
Using social media and crowdsourcing, Loh brought together a team, including many CMU students and alumni working in mathematics, human-computer interaction, app development and information security — all areas where CMU excels — to develop NOVID, which can anonymously and reliably trace people's exposure to COVID-19. It works by allowing users to self-report if they have tested positive for the coronavirus without requiring or recording any personal information. Instead, it periodically senses how close a user is to another anonymous person who also is using the app. By sensing these interactions, NOVID can use chains of contact to detect and alert someone if they previously encountered anyone who recently self-reported a positive case, or anyone who previously came into contact with someone who recently self-reported, while keeping all identities anonymous.
While a slew of contact tracing apps has popped up in various stages of development in response to the pandemic, Loh notes that NOVID could have significant advantages.
"As we look to move past self-isolating and sheltering-in-place, contact tracing will play an important role in getting people safely back to their normal activities," Loh said. "This can be automatically achieved by mobile apps, but only if they can accurately sense distance between people."
NOVID is the first major anonymous contact tracing app to use ultrasound technology to accurately gauge the distance between people. Many of the current apps use only Bluetooth radio waves to measure the signal strength between devices.
Bluetooth signal strength alone is insufficient for calculating distances. It is limited not only by the level of signal strength that different types of phones emit, but also by the variability imposed by whether the phone is in a pocket or purse, or blocked by a human body. By relying on signal strength, Bluetooth-based apps can only hope to calibrate distances through later analysis, and even that is still an unsolved problem.
Instead, NOVID augments Bluetooth with ultrasound, soundwaves with frequencies just outside the range of human hearing, which modern mobile phones can emit and record. The principle is the same as telling how far away a thunderstorm is by counting the seconds between a lightning strike and a thunderclap. Knowing that sound travels at approximately 343 meters per second, if you can measure the time it takes to transmit and receive an ultrasonic signal within several milliseconds of accuracy, you can use this measurement to determine distance within an accuracy of several feet. Because today's mobile phones have this capability, Loh leveraged this to allow NOVID to measure distances at the ranges most relevant to contact tracing.
At these close ranges, ultrasonic signals can also be detected through bags and pockets, and can go around obstacles by reflecting off ceilings without significantly affecting distance measurement.
A walkthrough of the NOVID app.
Drawing on his expertise in network theory and algorithms, Loh designed and implemented efficient graph-theoretic algorithms to ensure NOVID's servers could handle millions of users if they suddenly signed up.
In addition to being able to technologically support users, Loh knew he needed to design an app people would want to use. To do this, he reached out through his network, seeking talent that was deeply committed to fighting the COVID-19 pandemic.
The first to answer the call was senior Dean Dijour. Dijour worked closely with sophomore Bennett Huffman, both Carnegie Mellon students majoring in human-computer interaction and information systems, to develop the app's user-centered interface. Like a Fitbit allows users to track their fitness goals, NOVID allows users to follow their social interactions and see how well they are playing their part to stop the spread.
"The opportunity to use my skills to fight COVID-19 and return to normalcy is something I couldn't pass up," Dijour said.
Dijour and Huffman extensively (and virtually) tested dozens of app mockups with people of all ages before working together with a team, including many CMU alumni and students, to engineer the app for Android and iOS.
"I had an idea of how to help the world return to normalcy."
The team has been working full force for a month to bring their idea to fruition. Dijour said working with the team was an invaluable experience.
"We've all benefited in some way from CMU's strong support of research, innovation and interdisciplinary collaboration," he said.
With this unique approach, Loh was able to bring a polished product to the forefront of the technological response to the pandemic by combining the force of the CMU community and his social enterprise Expii to support the app's development.
The next step is to test the app in the real-world setting.
If the app proves effective, it could help employers as they consider new ways to keep their communities safe when returning to normal activities. Carnegie Mellon is considering how it might use the app in its plans to re-engage research on campus.
"As we consider options for return to work, we know that technologies like NOVID will play a key role in keeping our community safe and preventing a rebound outbreak," said J. Michael McQuade, vice president for research at CMU.
NOVID is currently available in beta version for Android devices at novid.org.