Freeway ramp control is traditionally based on the assumption that bottleneck capacity is constant and known a priori. This is certainly the case of the current Stratified Zone Metering (SZM) strategy implemented in the 260 miles freeway network of Minneapolis St. Paul metropolitan area. However, it is well known that capacity is not constant but varies with traffic conditions and is generally reduced when congestion sets in. To address this issue, an improved capacity estimation is presented here which is generally suitable for real time control and related applications such as simulation. Specifically using actual data from 41 detector stations over a one year period it is demonstrated that when bottleneck is uncongested, capacity is normally distributed and a recommendation is made on selecting the proper capacity value. When congestion sets in one can only talk about operational capacity which is generally lower and is a function of the congestion level. In such case, a moving average method based on real time measurements is proposed. The combined methodology is implemented for testing its effects on the Stratified Zone Metering (SZM) strategy through a state of the art micro-simulator. Results at two typical freeway sections suggest modest control strategy improvements. For example, System Total Travel Times and Freeway Delays reduced by as much as 5.45% and 7.46% respectively while Average System Speed increased by as much as 5.87%; Energy Consumption, Number of Stops and Pollution levels decreased by as much as 4.89%, 10.88%, 4.35% respectively.
Samenvatting