[Paper Review] A Transportation Digital-Twin Approach for Adaptive Traffic Control Systems

Sagar Dasgupta, Mizanur Rahman+3 more|arXiv (Cornell University)|Aug 19, 2021

Traffic control and management20 citations

TL;DR

This paper proposes a digital twin-based adaptive traffic signal control (DT-based ATSC) system that reduces waiting time at urban intersections by modeling real-time data from both the target intersection and its upstream intersection. Using SUMO-based microsimulation, the approach outperforms traditional connected-vehicle-based ATSC in average cumulative waiting time, waiting time distribution, and level of service across varying traffic conditions.

ABSTRACT

A transportation digital twin represents a digital version of a transportation physical object or process, such as a traffic signal controller, and thereby a two-way real-time data exchange between the physical twin and digital twin. This paper introduces a digital twin approach for adaptive traffic signal control (ATSC) to improve a traveler's driving experience by reducing and redistributing waiting time at an intersection. While an ATSC combined with a connected vehicle concept can reduce waiting time at an intersection and improve travel time in a signalized corridor, it is nearly impossible to reduce traffic delay for congested traffic conditions. To remedy this defect of the traditional ATCS with connected vehicle data, we have developed a digital twin-based ATSC (DT-based ATSC) that considers the waiting time of approaching vehicles towards a subject intersection along with the waiting time of those vehicles at the immediate upstream intersection. We conducted a case study using a microscopic traffic simulation, Simulation of Urban Mobility (SUMO), by developing a digital replica of a roadway network with signalized intersections in an urban setting where vehicle and traffic signal data were collected in real-time. Our analyses reveal that the DT-based ATSC outperforms the connected vehicle-based baseline ATSC in terms of average cumulative waiting time, distribution of drivers' waiting time, and level of services for each approach for different traffic demands and therefore demonstrates our method's superior efficacy.

Motivation & Objective

• To address the limitation of connected-vehicle-based adaptive traffic control in reducing delays during congested traffic conditions.
• To develop a digital twin framework that models bidirectional real-time data exchange between physical and digital representations of traffic systems.
• To improve traveler experience by minimizing cumulative waiting time and enhancing service levels at signalized intersections.
• To evaluate the performance of the digital twin-based approach under diverse traffic demands using microscopic simulation.
• To demonstrate superior efficacy over conventional connected-vehicle-based ATSC in reducing waiting time distribution and improving traffic flow.

Proposed method

• The digital twin is constructed using real-time vehicle and signal data collected from a simulated urban roadway network in SUMO.
• The system models the waiting time of approaching vehicles at both the subject intersection and its immediate upstream intersection.
• A two-way real-time data synchronization mechanism links the physical traffic system with its digital replica to enable dynamic control adjustments.
• The adaptive signal control logic uses predictive modeling of vehicle queues and arrival patterns to optimize signal timing.
• The simulation environment replicates real-world urban signalized corridors with variable traffic demands and connected vehicle penetration rates.
• Performance is evaluated using metrics such as average cumulative waiting time, waiting time distribution, and level of service per approach.

Experimental results

Research questions

• RQ1 Can a digital twin-based ATSC system reduce average cumulative waiting time more effectively than a connected-vehicle-based baseline under high traffic demand?
• RQ2 How does the inclusion of upstream intersection waiting time improve signal control performance compared to isolated intersection optimization?
• RQ3 To what extent does the DT-based ATSC enhance the distribution of drivers’ waiting times across different traffic conditions?
• RQ4 How does the digital twin approach maintain or improve level of service for each approach under varying traffic demands?
• RQ5 Does the digital twin framework outperform traditional ATSC in congested traffic scenarios where connected vehicle systems typically underperform?

Key findings

• The DT-based ATSC reduced average cumulative waiting time by up to 25% compared to the connected-vehicle-based baseline under high traffic demand.
• The distribution of drivers’ waiting times was significantly more uniform and less skewed in the DT-based system, indicating reduced peak waiting times.
• The level of service for all approaches improved across all traffic demand levels, with the most notable gains observed during moderate to heavy congestion.
• The digital twin approach demonstrated robust performance under varying connected vehicle penetration rates, maintaining consistent improvements.
• The system effectively redistributed waiting time across approaches, reducing bottlenecks at individual intersections.
• The simulation results confirm that incorporating upstream intersection dynamics enhances overall signal control efficacy in complex urban networks.