Metering Strategies Applied to Signalized Arterial Networks

This paper describes the development and adaptation of ramp metering strategies for use in signalized arterial networks. The strategies were developed in a simulation environment using the micro-simulation model NEMIS. The paper concentrates on the adaptation of the strategies for signalized arterials, producing real time control strategies which can deal with the problems of over-saturation and blocking back. The basic philosophy behind the strategies is to formulate the problem as a regulator problem, regulating the average link densities to a pre-specified level.

The NEMIS model is used to represent the real world, providing simulated detector loop data. This simulated loop data feeds another model developed by Cremer and Schoof, which estimates flows, queues, and link densities. This sub-model represents the UTC model which is present in systems such as SCOOT and SPOT, its only purpose being to introduce measurement errors similar to those which would occur in reality. These estimates are used to determine the control strategy for the immediate future, which is then enacted within NEMIS.

The paper presents detailed results for four algorithms with various parameters and draws conclusions about the effects of the parameters.