A Physics-Based Finite-State Abstraction for Traffic Congestion Control

TL;DR

This paper introduces a physics-based finite-state model for traffic congestion control in interconnected road networks, utilizing MPC and quadratic programming to optimize boundary inflow and traffic signals.

Contribution

It presents a novel finite-state abstraction of traffic networks combined with MPC for effective congestion control, integrating physics-based modeling with optimization techniques.

Findings

Traffic congestion can be effectively controlled through boundary inflow optimization.

Traffic signal phases can be optimized using receding horizon control.

Simulation results demonstrate successful congestion management.

Abstract

This paper offers a finite-state abstraction of traffic coordination and congestion in a network of interconnected roads (NOIR). By applying mass conservation, we model traffic coordination as a Markov process. Model Predictive Control (MPC) is applied to control traffic congestion through the boundary of the traffic network. The optimal boundary inflow is assigned as the solution of a constrained quadratic programming problem. Additionally, the movement phases commanded by traffic signals are determined using receding horizon optimization. In simulation, we show how traffic congestion can be successfully controlled through optimizing boundary inflow and movement phases at traffic network junctions.