Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Emission reduction potential of freeway stop-and-go wave smoothing

Junyi Ji, Derek Gloudemans, Gergely Zachár, William Barbour, Jonathan Sprinkle, Benedetto Piccoli, Daniel B. Work

Abstract

Real-world potential of stop-and-go wave smoothing at scale remains largely unquantified. Smoothing freeway traffic waves requires creating a gap so the wave can dissipate, but the gap suggested is often too large and impractical. We propose a counterfactual wave smoothing benchmark that reconstructs a smooth and feasible trajectory from each empirical trajectory by solving a quadratic program with fixed boundary conditions and a maximum allowable gap constraint. We estimate the emission reduction potential from trajectories using the MOVES model. Applying the framework to nine weeks of weekday peak traffic data, featuring rich day-to-day stop-and-go wave dynamics, from the I-24 MOTION testbed, we find meaningful reduction potential under a 0.1-mile maximum gap: average CO2 reductions of 7.92% to 12.04% across lanes, with concurrent reductions of 14.30% to 28.91% CO, 23.15% to 29.42% HC, and 24.37% to 30.98% NOx. Our analysis also quantifies the trade-off between maximum allowable gap opening and emissions benefits.

Emission reduction potential of freeway stop-and-go wave smoothing

Abstract

Real-world potential of stop-and-go wave smoothing at scale remains largely unquantified. Smoothing freeway traffic waves requires creating a gap so the wave can dissipate, but the gap suggested is often too large and impractical. We propose a counterfactual wave smoothing benchmark that reconstructs a smooth and feasible trajectory from each empirical trajectory by solving a quadratic program with fixed boundary conditions and a maximum allowable gap constraint. We estimate the emission reduction potential from trajectories using the MOVES model. Applying the framework to nine weeks of weekday peak traffic data, featuring rich day-to-day stop-and-go wave dynamics, from the I-24 MOTION testbed, we find meaningful reduction potential under a 0.1-mile maximum gap: average CO2 reductions of 7.92% to 12.04% across lanes, with concurrent reductions of 14.30% to 28.91% CO, 23.15% to 29.42% HC, and 24.37% to 30.98% NOx. Our analysis also quantifies the trade-off between maximum allowable gap opening and emissions benefits.
Paper Structure (21 sections, 7 equations, 13 figures, 1 table)

This paper contains 21 sections, 7 equations, 13 figures, 1 table.

Table of Contents

  1. Introduction
  2. Related work
  3. Stop-and-go wave smoothing strategies
  4. Emission reductions via wave smoothing
  5. Methodology
  6. Empirical trajectories
  7. Benchmarking smooth and feasible trajectories
  8. Emission model
  9. Data
  10. Results
  11. Case study on one empirical trajectory
  12. Large-scale analysis on one day
  13. Day-to-day emission reduction analysis
  14. Discussion
  15. Trade-off between the gap opening and emission reduction
  16. ...and 6 more sections

Figures (13)

  • Figure 1: Counterfactual wave smoothing benchmark comparison. Empirical trajectory example data dated from 2024-06-18, Lane 1, demonstrated under a 0.1-mile maximum gap constraint. (Top) Time-space diagram showing the original empirical trajectory and the benchmark trajectory constrained within the boundary range. (Bottom) Speed profiles highlighting the elimination of stop-and-go waves in the benchmark trajectory while maintaining consistent boundary conditions.
  • Figure 2: Overview of the study framework for benchmarking the emission costs of freeway stop-and-go waves: (a) Empirical vehicle trajectories from the I-24 MOTION testbed gloudemans202324, providing over two months of high-resolution data that captures rich stop-and-go wave dynamics; (b) Benchmarking scenarios is then proposed and smooth and feasible trajectories are reconstructed; (c) emission model as a function of speed and acceleration, we implement the U.S. EPA MOVES model USEPA_Motor_Vehicle_Emission_2024 for project-level estimation.
  • Figure 3: Operating Mode ID (OpModeID) map defined by speed and VSP, excluding braking (OpModeID 0) and idling (OpModeID 1), used in the latest U.S. EPA MOVES5 model USEPA_Motor_Vehicle_Emission_2024: each rectangle represents an operating mode with a unique ID, and further corresponds to a specific emission rate for different pollutants.
  • Figure 4: I-24 MOTION testbed. I-24 MOTION is a large-scale freeway instrument designed to observe freeway traffic congestions. It consists of 40 poles (as illustrated in this figure) with 276 high-resolution cameras covering a 4.2-mile stretch of I-24 near Nashville, TN.
  • Figure 5: Data used in this paper. Time-space diagrams for the data analyzed in this study collected from I-24 MOTION testbed. The data is collected from May 17, 2024 to July 19, 2024, covering 9 weeks of weekday peak-hour traffic (6:00 AM to 10:00 AM). Each day of data covers 4 hours of traffic over the 4.2-mile testbed, with rich stop-and-go wave dynamics. Holidays with light traffic and anomalous days with significant events (e.g. multiple lanes closed) are excluded from the analysis, as labeled with masks and * next to the dates.
  • ...and 8 more figures