A variable speed limit controller for recurrent congestion based on the optimal solution

Abstract

The main goal of this paper is the proposal and simulation of a SPEed limit controller for Recurrent Traffic jams (SPERT) that approximates the behavior of an optimal controller when congestion profiles are similar to the typical one. In order to achieve this goal, the optimal solution for the typical demand profile is computed and used as a first estimation for the logic-based controller. If the real congestion differs from the typical one, the values of the speed limits are adapted by advancing or delaying their activation and deactivation. Eleven scenarios have been considered in order to test to proposed controller under different traffic conditions. The results show that the proposed controller is able to approach the optimal behavior (with a better performance that previously proposed easy-to-implement VSL control algorithms) while eliminating on-line computational cost, and increasing robustness.