Integrating Simulation and Signal Processing with Stochastic Social Kinetic Model

Abstract

Data that continuously track the dynamics of large populations have spurred a surge in research on computational sustainability. However, coping with massive, noisy, unstructured, and disparate data streams is not easy. In this paper, we describe a particle filter algorithm that integrates signal processing and simulation modeling to study complex social systems using massive, noisy, unstructured data streams. This integration enables researchers to specify and track the dynamics of complex social systems by building a simulation model. To show the effectiveness of this algorithm, we infer city-scale traffic dynamics from the observed trajectories of a small number of probe vehicles uniformly sampled from the system. The experimental results show that our model can not only track and predict human mobility, but also explain how traffic is generated through the movements of individual vehicles. The algorithm and its application point to a new way of bringing together modelers and data miners to turn the real world into a living lab.