MobIntel: Sensing and analytics infrastructure for urban mobility intelligence

Mobility monitoring in urban environments provides valuable insights into pedestrian and vehicle movement. Understanding the causes and effects of changing mobility patterns can help city officials and businesses optimize operations and support economic development. In this paper, we introduce MobIntel, a privacy-centric alternative to visual surveillance for mobility monitoring. We present the design, implementation, and evaluation of the MobIntel sensor network, the associated cloud-based processing system, and the web-based visualization portal. We evaluate system capacity and accuracy based on real-life data collected from a deployment in downtown West Palm Beach, FL, spanning 14 months. We offer sample use cases where pedestrian activity patterns revealed by MobIntel can be helpful to the city government. Finally, we discuss several obstacles that must be overcome to transition from pedestrian counting to trajectory monitoring.