Traffic flow modeling and limitation on the coexistence of WAVE and WLAN

Traffic flow modeling is important to the performance analysis/evaluation of services provided by Vehicular Adhoc Networks (VANET), and is also a useful guidance to the deployment of VANET. Different from prior work based on empirical data collected decades ago, in this work, we collect a large amount of empirical traffic flow data from five typical overhead road segments during two different time periods recently in Shanghai. Statistical results in a short time scale (i.e., within half an hour) show that the lane-level traffic volumes/vehicles' velocities in the monitored road segments followed truncated Gaussian distribution (with a match rate of roughly 90%) better than Poisson distribution (with a match rate of roughly 80%) which is normally assumed in the literature. Traffic flow characteristic in a long time scale (i.e., in a day or a week) is also presented, which shows that the traffic density in the night and daytime are very different, however, the traffic flow density in the daytime stays high. With the obtained traffic flow characteristics, we discuss the possibility of the coexistence of WAVE and WLAN in 5.9G band according to the Federal Communications Commission's intention, and point out that for areas in the vicinity of a overhead road inside Outer Ring road in Shanghai, it is not practical for WLAN devices to operate on 5.9G WAVE band.