Modeling virus transmission risks in commuting with emerging mobility services: A case study of COVID-19

Commuting is an important part of daily life. With the gradual recovery from COVID-19 and more people returning to work from the office, the transmission of COVID-19 during commuting becomes a concern. Recent emerging mobility services (such as ride-hailing and bike-sharing) further deteriorate the infection risks due to shared vehicles or spaces during travel. Hence, it is important to quantify the infection risks in commuting. This paper proposes a probabilistic framework to estimate the risk of infection during an individual's commute considering different travel modes, including public transit, ride-share, bike, and walking. The objective is to evaluate the probability of infection as well as the estimation errors (i.e., uncertainty quantification) given the origin-destination (OD), departure time, and travel mode. We first define a general trip planning function to generate trip trajectories and probabilities of choosing different paths according to the OD, departure time, and travel mode. Then, we consider two channels of infections: 1) infection by close contact and 2) infection by touching surfaces. The infection risks are calculated on a trip segment basis. Different sources of data (such as smart card data, travel surveys, and population data) are used to estimate the potential interactions between the individual and the infectious environment. A first-order approximation is used to simplify the computational complexity. We also derive the closed-form formulation for the estimation errors, enabling us to quantify the uncertainty of the estimation results. The model is implemented in the MIT community as a case study. We evaluate the commute infection risks for employees and students. Results show that most of the individuals have an infection probability close to zero. The maximum infection probability is around 1.3%, implying that the probability of getting infected during the commuting process is low. Individuals with larger travel distances, traveling in transit, and traveling during peak hours are more likely to get infected. Practical implementations of the model are also discussed.