The ‘Taxi Driver’ Process and Passive Mobile Sensing on Networks

Abstract

The modern world offers many examples and potential applications of mobile sensing agents passively traversing networks. However, the dynamics of these systems and their potential sensing power is relatively unexplored. In this paper, I examine a model designed to simulate this kind of process for urban sensing, extending it beyond urban street networks to general graphs. I calculate the number of visits, the mean first passage time, and the expected return time in the simplest case, and provide a bound and a scaling for the number of visits in more complex cases. Finally, I examine the spatial predictiveness of the model on a dataset of New York City taxi trips.