A System Architecture to Aggregate Video Surveillance Data in Smart Cities

In recent years, we have experienced many initiatives in developing smart cities across the world. In these densely populated cities, video surveillance will play an integral role to ensure the safety of the citizens. The major challenge will be to transport high volumes of data generated from dense deployment of video cameras throughout the city to a central aggregation facility for analysis and storage. To address this problem, we propose a system solution using existing public bus transit system to collect data from the cameras and physically transport it to the bus terminus, to be uploaded to the data center. The system uses heterogeneous wireless network interfaces, where the camera nodes form a mesh network to route the data to the nearest bus stop for offloading. We utilize high capacity links using mmWave devices at bus stops to offload the data to the buses when they make routine stops. We also propose a novel routing protocol, R2H, to route the data generated from these cameras to a bus stop while minimizing the time to offload the video data to an incoming bus. The protocol is designed to be aware of the various challenges that are specific to this system like bus schedule and video payload. We have evaluated our system using wireless experiments and simulation using actual bus route maps and times from New York city area. Our analysis shows that R2H achieves 60% less end to end delay compared to OLSR under varying video payload and time of the day.