Exploring Urban Dynamics Based on Pervasive Sensing: Correlation Analysis of Traffic Density and Air Quality

Modern cities, with large population and complicated infrastructures, are complex entities with non-linear and dynamic properties that challenge the city management. Therefore, as the first step towards the goal of thorough understanding of the phenomena, pervasive urban sensing have become a cornerstone of future smart city that enhance the interplay between the cyber space and the physical world. We introduce a taxi-based pervasive urban sensing system and its key algorithm, aiming at the quantitative study of the correlation between human activities and environmental changes. Our contributions are twofold. First, we propose an urban crowd-sourcing framework that take automobiles as participatory mobile agents to the sensing tasks, and implemented a prototype in Beijing. Second, we design a Spatial-Temporal Manifold Learning (STML) algorithm to analyse the correlation between physical processes. Based on noisy and partially labelled dataset that are collected by pervasive urban sensor networks, we evaluate STML's performance by analysing correlation between the traffic density and the air quality. The results show great potential of STML for future urban sensing applications.