The Selective Random Subspace Predictor for Traffic Flow Forecasting

In this paper the authors describe a coordinated ramp control system that is incident-responsive and based on a stochastic optimal control approach. The methodology proposed has two major components. The first is a generalization of sequential probability ratio testing technology to be applied to congestion patterns in control zones. The second is a coupling of extended Kalman filtering with stochastic optimal control for use in group-based ramp control. Authors have found that this approach is useful for responding to dynamic traffic congestion conditions caused by traffic incidents. The results of the authors’ method are compared to the three types of control methods currently in place at the site in question. The authors find the results of the method in question to be more advantageous in easing traffic congestion than those of the methods currently in place.

  • Availability:
  • Authors:
    • Sun, Shiliang
    • Zhang, Changshui
  • Publication Date: 2007-6


  • English

Media Info

Subject/Index Terms

Filing Info

  • Accession Number: 01054576
  • Record Type: Publication
  • Source Agency: UC Berkeley Transportation Library
  • Files: BTRIS, TRIS
  • Created Date: Jun 18 2007 8:32PM