An extended model predictive control approach to coordinated ramp metering

• Published in: 2013 10th IEEE International Conference on Control and Automation (ICCA) pp. 681 - 686
• Main Authors: Linglong Hu, Weili Sun, Hui Wang
• Format: Conference Proceeding
• Language: English
• Published: IEEE 01.06.2013
• Subjects: Optimal control; Optimization; Predictive control; Predictive models; Traffic control; Trajectory; Vehicles
• ISBN: 9781467347075; 1467347078
• ISSN: 1948-3449
• DOI: 10.1109/ICCA.2013.6564996

Ramp metering is the most powerful tool to reduce congestion in freeway management. The problem of designing an on-line coordinated ramp metering strategy is considered in this paper. First, the prevalent model predictive control (MPC) structure is used to formulate the problem with a second-order macroscopic traffic flow prediction model. To reduce the computational effort to match the request of on-line application, the original MPC is then extended with a controller positioning procedure. Not all but only the necessary ramps are selected as controllers by investigating the partial derivation of the cost criterion with respect to each ramp metering rate. Finally, a gradient-based numerical solving algorithm is introduced and the overall method is validated in a hypothetic network. Applying the proposed method, the performance of the freeway network improves as much as 13.4% compared with no-control case. In comparison with the original MPC method, the extended MPC method saves 32.3% computation time while barely degrades the control effectiveness.