RAPID: Acoustic Communications and Sensing for COVID-19 Data Collection

The outbreak of the novel coronavirus (COVID-19) has unfolded as a major international crisis whose influence has extended to every aspect of daily life. This has led to a devastating loss of human life as well as the nation's economic well-being. As various US states begin to restart their economies, contact tracing has become an invaluable tool in allowing the workforce to return in a safe and controlled manner. The proposed project aims to build a privacy-preserving crowdsensing system for effective COVID-19 tracking and tracing by leveraging the ubiquity of mobile devices. This work overcomes key challenges of infrastructure-based techniques such as video monitoring systems, etc., that are difficult to scale and provide broad coverage. This approach uses a unique inaudible acoustic based communication system to identify and monitor persons with whom users have interacted with. It uses only acoustic signal transmission with every day mobile phone sensors (i.e., inbuilt speakers and microphones) to facilitate easy implementation. The research conducted during this project will overcome key challenges in energy management and power control to prolong battery life, dealing with an occluded environment where the phone may be in a pocket or a purse, and detecting physical interactions between users, thus making fundamental contributions to domain specific research areas themselves.

The system will leverage the normal procedure of human interactions in the context of social encounters and adapt a novel acoustic signals dissemination service that selectively broadcasts information within particular "turfs". The encounters' information will be uploaded to a central server either automatically or with users' manual processing. The map at the central server will be updated accordingly. The solution will preserve a number of desirable qualities in the following dimensions: (i) The sensing range of acoustic signals on mobile phones is small compared with competing systems such as Bluetooth, but meets the requirements on virus detection. In fact, this feature enables far less false positives than Bluetooth based approaches. The short range also helps with privacy. (ii) Each user's unique ID such as WiFi and Bluetooth MAC addresses would not be disclosed to their peer encounters. Instead, only randomly generated IDs are disclosed. (iii) Users can choose what information to report to the central server, e.g., their encounters with/without GPS, their medical condition, age, real ID, etc. They could also choose not to report certain encounters or encounters at certain locations.

This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.