Edge Sensor and Machine Learning Based Ventilation Management to Reduce Airborne Disease Spread

Danika Gupta, Meenakshi Nair, Shriya Shukla

Abstract: COVID-19 has impacted billions of people and devastated economies around the world. We are now at the stage of the pandemic where businesses need to stay open even though COVID-19 is still present in the community. In this paper, we explore the relationship between disease spread and ventilation, and explore approaches for intelligent ventilation management. Using public datasets and ad-hoc experimentation, we demonstrate how a combination of edge sensors and machine learning can be used effectively to monitor and forecast CO2 levels. Ad-hoc experiments show the wide range of levels possible in public spaces. Next, using air quality measurements from the Mall of Tripla, we demonstrate the efficacy of machine learning algorithms to forecast CO2 and also predict it from other environmental data. Finally we describe our mobile app that enables business operators to leverage CO2 measurements to improve ventilation and COVID-19 safety.