Interactive Physics-Inspired Traffic Congestion Management

TL;DR

This paper introduces a physics-inspired, model predictive control approach for managing traffic congestion in interconnected road networks by modeling traffic flow dynamics and optimizing boundary controls.

Contribution

It presents a novel physics-based framework combining mass conservation, diffusion dynamics, and heat flux analogy for traffic management, with interactive boundary control strategies.

Findings

Successfully manages congestion in large, heterogeneous NOIR simulations.

Demonstrates effectiveness of model predictive boundary control.

Integrates heat flux analogy into traffic flow modeling.

Abstract

This paper proposes a new physics-based approach to effectively control congestion in a network of interconnected roads (NOIR). The paper integrates mass flow conservation and diffusion-based dynamics to model traffic coordination in a NOIR. The mass conservation law is used to model the traffic density dynamics across the NOIR while the diffusion law is applied to include traffic speed and motion direction into planning. This paper offers an analogy between traffic coordination in a transportation system and heat flux in a thermal system to define a potential filed over the NOIR. The paper also develops an interactive light-based and boundary control to manage traffic congestion through optimizing the traffic signal operations and controlling traffic flows at the NOIR boundary nodes. More specifically, a model predictive boundary control optimizes the NOIR inflow traffic while a receding horizon optimizer assigns the optimal movement phases at the NOIR intersections. For simulation, the paper models traffic congestion in a heterogeneous NOIR with a large number of interior and boundary nodes where the proposed interactive control can successfully manage the congestion.