Algorithm for Controlling Spillback from Ramp Meters

Abstract

The queue control algorithms that are commonly used by ramp-metering controllers often permit the queue to extend a considerable distance upstream of the queue detector on the ramp. This detector senses the presence of a queue with the intent of preventing queue spillback onto the surface street or intersection that is located upstream of the ramp. Queue control is usually obtained by switching between the desired metering rate and a higher metering rate. In some cases metering is terminated. This type of limit cycle control often results in considerable variation in the length of the queue, even for a constant average vehicle arrival rate at the ramp. This high-amplitude limit cycle operation is indicated, by simulation, to result from the relatively long detector sampling and computation period (typically 1 min) that is commonly used. A control algorithm using a 10-sec sampling period is indicated to substantially reduce the limit cycle amplitude. Performance is further improved when the control algorithm anticipates the buildup of the queue over the spillback detector.

Coordinated Ramp-Metering Control Using a Time-Gap Based Traffic Model

“…On-ramp queue management is used to prevent on-ramp queue spillover onto neighboring arterial streets by increasing ramp-metering rate [10] [11]. This strategy can have an adverse effect on highway mainline traffic and thus diminish the benefits of ramp-meter control.…”

Freeway Traffic Density and On-Ramp Queue Control via ILC Approach

“…When the queue detector occupancy falls below the threshold, the metering rate is reset to the value determined by the controller. It has been noted [19,20] that this queue-override scheme leads to oscillatory behavior and underutilization of on-ramp storage capacities.…”

A localized switching ramp-metering controller with a queue length regulator for congested freeways

“…It can be easily shown that the resulting closed loop dynamics is not asymptotically stable, given that the open loop queue length dynamics is that of a simple integrator. It has been noted (Gordon, 1996;Smaragdis and Papageorgiou, 2004) that this queue-override scheme leads to oscillatory behavior and under-utilization of onramp storage capacities. Gordon (1996) attempted to improve the performance of the queue-override scheme by filtering the occupancy signal and reducing the sampling time interval.…”

“…It has been noted (Gordon, 1996;Smaragdis and Papageorgiou, 2004) that this queue-override scheme leads to oscillatory behavior and under-utilization of onramp storage capacities. Gordon (1996) attempted to improve the performance of the queue-override scheme by filtering the occupancy signal and reducing the sampling time interval. Smaragdis and Papageorgiou (2004) proposed a proportional controller that relies on the on-ramp vehicle demands.…”