Model predictive control for ramp metering combined with extended Kalman filter-based traffic state estimation.

Abstract

Ramp metering is a dynamic traffic control measure that has proven to be very effective. There are several methods to determine appropriate ramp metering signals for a given traffic situation. In this paper, a framework consisting of model predictive control (MPC) for ramp metering, combined with extended Kalman filter-based (EKF) traffic state estimation is presented. Based on traffic measurements at a limited number of locations, the EKF is able to provide the MPC ramp metering controller with estimations of the traffic states in the motorway segments of the motorway stretch under control. By using the same traffic flow model in the EKF as in the MPC prediction model, some important model parameters of the MPC prediction model can be estimated and be fed directly to the MPC controller. This functionality enables the MPC prediction model to track changes in the traffic system (e.g. due to weather conditions, incidents, etc.). The presented EKF-MPC controller for ramp metering is simulated for a case study on the E17 motorway Ghent–Antwerp in Belgium.