Towards a Practical Application of Model Predictive Control to Suppress Shock Waves on Freeways

We present the results of the application of model predictive control (MPC) to a micro-simulation model with a scenario where shock waves are present, and a microsimulation model functions as a substitute for the real-world traffic system. Shock waves emerge in most cases from traffic jams at bottlenecks, propagate upstream on the freeway, and can remain existent for a long time and distance. This increases travel time, is potentially unsafe, and increases noise and air pollution.

Previously reported results using MPC to eliminate shock waves, showed an improvement of 20% of the total time that the vehicles spent in the network. However, they were based on the assumption that the simulation model (representing the real world) and the prediction model are the same, which may have lead to overoptimistic results.

In this paper a micro-simulation model (Paramics 5.1 by Quadstone) is used to represent the real world, which results in a model mismatch between the simulation model and the prediction model. We show by simulation that even in the case of a model mismatch the controller is able to suppress or remove shock waves.