Project
| # | Title | Team Members | TA | Documents | Sponsor |
|---|---|---|---|---|---|
| 64 | CHEAP, ACCURATE, AND PRIVACY-PRESERVING CONTACT TRACING CHIP |
Abhinav Singh Anshul Sanamvenkata Kapil Kanwar |
Ali Kourani | design_document1.pdf design_document2.pdf final_paper1.pdf other1.pdf |
|
| # Team members - Anshul Sanamvenkata (anshul2), In Person - Abhinav Singh (abhinav6), In Person - Kapil Kanwar (kkanwar2), Online # Problem COVID is a deadly and highly infectious disease, and given the current trend of globalization and environmental destruction, such pandemics will only become more common. Testing and contact tracing are one of the best ways to fight a highly infectious disease while allowing people to maintain some semblance of a normal life. Current contact tracing solutions either rely on manual effort, or mobile apps, which are both flawed. Manual methods typically involve calling someone who has tested positive and asking them to recall whom they met, which obviously is highly imperfect. Mobile apps typically track people's locations to look for potential contacts, which raises privacy concerns, not to mention that GPS is not always available, or particularly granular. While there are mobile app solutions that use NFC, bluetooth, or even audio, these all have flaws, and additionally are not possible on all phones, not to mention that there are parts of the world in which modern smartphones are not widely available to begin with due to high cost. # Solution Overview We propose a small, cheap chip that can be easily carried which will automatically communicate with other nearby chips to perform contact tracing. Each chip is assigned an anonymous ID and communicates its ID and COVID status with any nearby chips to determine the edges of a contact graph. To identify potential transmission, we require that the chip be connected to a low-end smartphone or PC to upload its anonymized contact data to a server at some regular interval (~once a day), which will allow the server to construct a fully anonymous contact graph and notify users of possible transmission in a timely fashion. Privacy is ensured in the following manner: testing centers provide the user with a cryptographically signed COVID status message (positive or negative) concatenated with the user ID to prevent replay attacks, and the signed message is uploaded to the server along with the user’s anonymous ID. This way, the server is certain the user is not lying about their COVID status, but the server also doesn’t know who the user is. Why is this better than existing solutions? We can very accurately determine distances using ultra wideband, addressing the flaws of NFC, bluetooth, and audio, which are not accurate enough and simply rely on the fact that signals get weaker over distance. By virtue of this not requiring the latest smartphones with the latest software / hardware, we also address the cost issues. Finally, since we only send anonymous contact information to the server, and no location or personal information, there are virtually no privacy issues. # Solution Components - **Communication**: To communicate between contact tracing chips we decided on using Ultra Wideband technology which is a communication protocol similar to Bluetooth. The benefit is that it is much easier to get proximity information between different devices using this protocol. The distance between two UWB devices can be measured precisely by measuring the time that it takes for a signal to perform a round-trip between the two devices (with some fixed latency offset), while Bluetooth is much less accurate. - **Microcontroller**: This subsystem acts as the main control system for the project. It communicates with the Ultra Wideband chip through SPI communication and takes care of the necessary handshake between them, calculating the proximity between them, and verifying / storing COVID statuses from other devices. - **Power**: While the goal of this project is to make the system as low energy as possible, it is inevitable that it will need to be recharged and we plan on using an array of small rechargeable coin cell batteries and a simple Li-Po balance charger circuit that uses USB to charge it. (One of our group members has experience with designing a Li-Po balance charger so we know it is possible) - **Docking**: Consists of a program and device driver that can be run on a PC that will communicate with the chip over USB and with the server to download and upload contact data. Additionally, to fully complete this subsystem, the chip’s microcontroller must also be programmed to communicate with the device driver over USB. - **Server**: A simple API to receive contact information (essentially edges in the contact graph), and let users know if they may have been in contact with someone who tested positive # Criteria for Success - Able to accurately transmit contact tracing data over short distances and reduce false positives by eliminating “contacts” that are significantly more than 6 feet apart - Maintains user privacy through cryptographic methods - neither the user’s location nor identity should be deducible using the data on the server - Must be cheap and low-power, to serve as an effective option in places where smartphones are not available |
|||||