Application of Sensor-Cloud Systems: Smart Traffic Control

Smart transportation paradigm has been treated as a feasible solution to ease the pressures caused by the rapid growth of motor vehicles in the urban area. As a key building block, smart traffic signal control has motivated many efforts in both academia and industry due to its promised gains. State-of-the-art proposals rely heavily on a powerful centralized computation infrastructure to handle huge amount of heterogeneous traffic data gathered by diversified sensors and actuators. However, this process will typically incur very large response latency, which is also the main barrier for their real world deployment. To realize near real-time traffic signal control, traffic data need to be processed at the "edge" (i.e. the generated position). Hence, we in this paper propose a fog computing based traffic signal control architecture, in which the phase timing task for a single intersection will be handled by a local fog node in a timely fashion, and global or regional optimization task will be left for the centralized cloud. In this manner, a tradeoff between local optimization and global optimization can be achieved. Moreover, we address the challenges and open research problems of the proposed architecture in hope to provide insights and research directions for modern traffic control.