Vehicle-infrastructure cooperation via the Connected Vehicle initiative is a promising mobile data source for improving real-time traffic management applications such as adaptive signal control. This paper focuses on developing estimation methods with the use of Connected Vehicle data for several measures of effectiveness (e.g., queue length, average speed, number of stops), essential for determining traffic conditions on urban signalized arterials for real-time applications. This research systematically determines minimum penetration rates that allow accurate estimates for a wide range of measures of effectiveness in undersaturated traffic conditions. The estimation of these measures and minimum penetration requirements has been tested using Next Generation Simulation (NGSIM) data.
This paper studies how Connected Vehicle (CV) routing can be leveraged to tame the propagation of congestion after the onset of an incident. To do so, the study focuses on an idealized one-way grid network with a homogenous demand pattern. Congestion propagation dynamics is first modeled analytically and simulated in Aimsun considering a scenario with non-adaptive agents. The results are then compared with a set of scenarios in which a percentage of the actors are CVs who communicate with a centralizer server that reports real-time information on the network state. This information is considered by the vehicles to expand the historic path choice set and repeatedly update their routing choices. The results show that CVs can drastically decrease congestion and improve its dissipation, even at low penetration rate values. The study shows that the CV technology in combination with collaborative routing algorithms can bring substantial benefits to urban mobility in a context of spatial scarcity for the private vehicle.