Suppression of Oscillations in Two-Class Traffic by Full-State Feedback

Abstract

This paper develops a full-state feedback controller that damps out oscillations in traffic density and traffic velocity whose dynamical behavior is governed by the linearized two-class Aw-Rascle (AR) model. Thereby, the traffic is considered to be in the congested regime and subdivided in two classes whereas each class represents vehicles with the same size and driver’s behavior. The macroscopic second-order two-class AR model consists of four first order hyperbolic partial differential equations (PDEs) and introduces a concept of area occupancy to depict the mixed density of two-class vehicles in the traffic. Moreover, the linearized model equations show heterodirectional behavior with both positive and negative characteristic speeds in the congested regime. The control objective is to achieve convergence to a constant equilibrium in finite time. The control input is realized by ramp metering acting at the outlet of the considered track section. The backstepping method is employed to design full-state feedback for the 4 × 4 hyperbolic PDEs. The performance of the full-state feedback controller is verified by simulation.