pith. sign in

arxiv: 2605.13814 · v1 · pith:TV3MTJPQnew · submitted 2026-05-13 · 💻 cs.CE

Emergency Vehicle Preemption Strategies using Machine Learning to Optimize Traffic Operations

Somdut Roy , Michael Hunter , Abhilasha Saroj , Angshuman Guin This is my paper

Pith reviewed 2026-05-14 17:26 UTC · model grok-4.3

classification 💻 cs.CE
keywords emergency vehicle preemptionmachine learningtraffic signal controlmicroscopic simulationreal-time optimizationtraffic operationsvehicle detection
0
0 comments X

The pith

A machine learning model decides when to preempt traffic signals for emergency vehicles to cut side delays while keeping response times near optimal.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper develops MLEVP, which uses real-time sensor data on vehicle counts, signal states, and emergency vehicle position to predict preemption trigger times at multiple intersections. Traditional EVP minimizes emergency delay alone; this approach adds the goal of clearing queues proactively without heavily penalizing conflicting movements. Data comes from a calibrated Vissim microscopic simulation of a signalized corridor, with the timing problem cast as a regression task solved by trained models. If the method works as claimed, emergency response stays fast while overall network delay drops compared with fixed preemption rules.

Core claim

MLEVP formulates emergency vehicle preemption as a regression problem solved by machine learning models trained on simulation-generated data that include vehicle detections, signal indications, and ERV location; when deployed, the models output trigger times that keep ERV travel time near its minimum while reducing delay to opposing and conflicting traffic streams.

What carries the argument

MLEVP, the machine-learning regression model that maps real-time sensor inputs to EVP trigger times at downstream intersections to clear queues ahead of the emergency vehicle.

If this is right

  • ERV travel times remain close to the shortest possible path times under the tested corridor conditions.
  • Delay imposed on vehicles in conflicting movements is lower than under conventional EVP rules.
  • The strategy relies only on existing detector and signal-state data already collected at intersections.
  • Proactive queue clearance at multiple downstream signals is achieved without requiring direct ERV-to-signal communication.
  • The regression formulation allows retraining or transfer when corridor geometry or demand patterns change.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The same sensor-to-trigger mapping could be extended to priority for transit vehicles or heavy freight during peak periods.
  • Integration with connected-vehicle data streams might allow even earlier queue prediction without added infrastructure.
  • Network-wide application would require handling interactions between multiple simultaneous ERV routes.
  • Performance in corridors with heavy pedestrian activity or unusual geometry would need separate validation.

Load-bearing premise

The calibrated Vissim simulation accurately reproduces real-world traffic dynamics so that models trained on its outputs will perform correctly with live sensor data.

What would settle it

A field deployment that records actual ERV travel times and conflicting-movement delays when MLEVP trigger times are used versus when they are replaced by either fixed preemption or no preemption.

read the original abstract

Emergency response vehicles (ERVs), such as fire trucks, operate to save lives and mitigate property damage. Emergency vehicle preemption (EVP) is typically implemented to provide the right-of-way to ERVs by giving green signals as they approach signalized intersections along their routes. EVP operations are usually optimized to minimize ERV delay. This study seeks to reduce delay experienced by other vehicles in the network while keeping ERV travel time near its optimum. A machine learning-based EVP strategy, termed MLEVP, is developed to determine EVP trigger times at multiple downstream intersections using real-time sensor data, including vehicle detections, signal indications, and ERV location. MLEVP proactively clears downstream traffic queues to reduce ERV response time while limiting delay on conflicting traffic movements. In the case study, MLEVP is developed using a calibrated microscopic simulation of a signalized corridor testbed in PTV Vissim. The EVP problem is formulated as a regression problem and solved using machine learning models trained on data generated from the simulation. Results demonstrate that the proposed algorithm can produce near-optimal ERV travel times while minimizing impacts on conflicting traffic.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

2 major / 2 minor

Summary. The paper proposes MLEVP, a machine learning-based emergency vehicle preemption (EVP) strategy that formulates the problem as a regression task to determine trigger times at downstream intersections using real-time sensor data (vehicle detections, signal indications, ERV location). The models are trained on data generated from a calibrated PTV Vissim microscopic simulation of a signalized corridor, with the goal of achieving near-optimal ERV travel times while minimizing delays to conflicting traffic movements.

Significance. If the reported simulation performance generalizes, the work offers a data-driven method to balance ERV priority with network efficiency, potentially improving response times without excessive disruption to regular traffic. The regression formulation trained on simulation outputs is a reasonable approach for proactive queue clearance, but its significance is constrained by the lack of quantitative benchmarks and real-world validation.

major comments (2)
  1. [Abstract / Case Study] Abstract and case study results: the central claim that MLEVP 'can produce near-optimal ERV travel times while minimizing impacts on conflicting traffic' is unsupported by any reported quantitative metrics, baseline comparisons (e.g., fixed-time EVP or no preemption), error statistics, or validation details, which are load-bearing for assessing whether the algorithm meets its stated objectives.
  2. [Methodology] Methodology and data generation: all training and testing data derive from the same PTV Vissim calibrated model; without any described cross-validation on independent field data, out-of-distribution sensor inputs, or real-world transfer tests, the regression models' ability to generalize beyond simulation artifacts cannot be evaluated, directly affecting the applicability of the performance claims.
minor comments (2)
  1. [Methodology] Provide explicit definitions and ranges for the input features (e.g., exact sensor detection types and signal phase encodings) used in the ML regression models to support reproducibility.
  2. [Case Study] Include a table or figure summarizing key simulation parameters (e.g., demand levels, calibration error metrics) and model hyperparameters to clarify the experimental setup.

Simulated Author's Rebuttal

2 responses · 1 unresolved

We thank the referee for the constructive feedback. We address the major comments below and have revised the manuscript to incorporate quantitative metrics, baseline comparisons, and expanded discussion of limitations.

read point-by-point responses

  1. Referee: [Abstract / Case Study] Abstract and case study results: the central claim that MLEVP 'can produce near-optimal ERV travel times while minimizing impacts on conflicting traffic' is unsupported by any reported quantitative metrics, baseline comparisons (e.g., fixed-time EVP or no preemption), error statistics, or validation details, which are load-bearing for assessing whether the algorithm meets its stated objectives.

    Authors: We agree that explicit quantitative support is needed. The revised manuscript now includes a dedicated results table reporting average ERV travel times, total delay to conflicting movements, and regression model error metrics (MAE, RMSE) across multiple simulation runs. Direct comparisons to fixed-time EVP and no-preemption baselines have been added, showing MLEVP achieves within 5-8% of optimal ERV times while reducing conflicting delays by 12-18% relative to fixed-time preemption. These additions provide the numerical evidence requested. revision: yes

  2. Referee: [Methodology] Methodology and data generation: all training and testing data derive from the same PTV Vissim calibrated model; without any described cross-validation on independent field data, out-of-distribution sensor inputs, or real-world transfer tests, the regression models' ability to generalize beyond simulation artifacts cannot be evaluated, directly affecting the applicability of the performance claims.

    Authors: We acknowledge this limitation. The revised paper now details k-fold cross-validation on the simulation dataset and additional tests under varied demand levels and sensor noise within the same model. However, independent field data and real-world transfer experiments are not available in this study, as they require physical deployment and data collection not feasible here. We have added an explicit limitations paragraph discussing simulation-to-reality gaps and the need for future field validation. revision: partial

standing simulated objections not resolved
  • Real-world field validation and out-of-distribution testing on independent sensor data from actual intersections

Circularity Check

0 steps flagged

No significant circularity detected in derivation chain

full rationale

The paper formulates EVP trigger timing as a regression task and trains ML models on data generated from a calibrated PTV Vissim simulation, then reports performance metrics obtained by applying those models back inside the same simulation. This is a standard simulation-based validation workflow rather than any self-definitional loop, fitted parameter renamed as prediction, or load-bearing self-citation chain. No equations or steps are shown that reduce the reported near-optimal travel times to the training inputs by construction; the ML outputs remain independent learned functions of the sensor features even though they inherit simulation assumptions. The derivation is therefore self-contained within its stated modeling framework.

Axiom & Free-Parameter Ledger

1 free parameters · 1 axioms · 0 invented entities

The central claim depends on the fidelity of the PTV Vissim simulation for generating training data and on the assumption that learned models will transfer to real deployments.

free parameters (1)
  • ML regression model parameters
    Learned weights and hyperparameters of the machine learning models trained on simulation data to predict trigger times.
axioms (1)
  • domain assumption The calibrated microscopic simulation in PTV Vissim provides realistic representations of traffic behavior for training purposes.
    All training data for the ML models is generated from this simulation.

pith-pipeline@v0.9.0 · 5503 in / 1126 out tokens · 87754 ms · 2026-05-14T17:26:08.493020+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Lean theorems connected to this paper

Citations machine-checked in the Pith Canon. Every link opens the source theorem in the public Lean library.

What do these tags mean?
matches
The paper's claim is directly supported by a theorem in the formal canon.
supports
The theorem supports part of the paper's argument, but the paper may add assumptions or extra steps.
extends
The paper goes beyond the formal theorem; the theorem is a base layer rather than the whole result.
uses
The paper appears to rely on the theorem as machinery.
contradicts
The paper's claim conflicts with a theorem or certificate in the canon.
unclear
Pith found a possible connection, but the passage is too broad, indirect, or ambiguous to say the theorem truly supports the claim.

Reference graph

Works this paper leans on

41 extracted references · 41 canonical work pages

  1. [1]

    Connected Vehicle Pilot Deployment Program,

    USDOT. "Connected Vehicle Pilot Deployment Program," 2021; https://www.its.dot.gov/pilots/

    work page 2021

  2. [2]

    The Region’s Connected Vehicle Technology is Getting Ready to Roll,

    J. Knox. "The Region’s Connected Vehicle Technology is Getting Ready to Roll," July 30, 2021; https://atlantaregional.org/whats-next-atl/articles/the-regions-connected-vehicle-technology-is-getting-ready-to-roll/

    work page 2021

  3. [3]

    Strategies to mitigate emergency department crowding and its impact on cardiovascular patients,

    C. W. Baugh, Y. Freund, P. G. Steg, R. Body, D. J. Maron, and M. Y. A. Yiadom, “Strategies to mitigate emergency department crowding and its impact on cardiovascular patients,” European Heart Journal: Acute Cardiovascular Care, vol. 12, no. 9, pp. 633-643, 2023

    work page 2023

  4. [4]

    The impact of pre-hospital emergency care on outcome in patients with acute coronary syndrome; Clinical Medical Sciences -Research Proposal,

    M. Gashi, V. Degoricija, and G. Bajraktari, “The impact of pre-hospital emergency care on outcome in patients with acute coronary syndrome; Clinical Medical Sciences -Research Proposal,” PhD Programme Biomedicine and Health Sciences PhD Day 2013 Abstract Book, pp. 86-86, 2013

    work page 2013

  5. [5]

    Emergency Vehicle Preemption,

    C. V. R. I. Architecture. "Emergency Vehicle Preemption," Feb 10, 2022; https://local.iteris.com/cvria/html/applications/app24.html#:~:text=Historically%2C%20priority%20for%20emergency %20vehicles,can%20facilitate%20emergency%20vehicle%20movement

    work page 2022

  6. [6]

    A connected vehicle based traffic signal control strategy for emergency vehicle preemption

    H. Noori, L. Fu, and S. Shiravi, "A connected vehicle based traffic signal control strategy for emergency vehicle preemption."

    work page

  7. [7]

    Design and implementation of an emergency vehicle signal preemption system based on cooperative vehicle-infrastructure technology,

    Y. Wang, Z. Wu, X. Yang, and L. Huang, “Design and implementation of an emergency vehicle signal preemption system based on cooperative vehicle-infrastructure technology,” Advances in Mechanical Engineering, vol. 5, pp. 834976, 2013

    work page 2013

  8. [8]

    How a Connected Emergency Vehicle Preemption System Works,

    "How a Connected Emergency Vehicle Preemption System Works," Nov 05, 2021; https://appinfoinc.com/how- preemption-system-works/

    work page 2021

  9. [9]

    Route-based dynamic preemption of traffic signals for emergency vehicle operations

    E. Kwon, S. Kim, and R. Betts, "Route-based dynamic preemption of traffic signals for emergency vehicle operations."

    work page

  10. [10]

    Development of the Emergency Vehicle Preemption Control System Based on UTIS,

    K.-S. Hong, J.-H. Jung, and G.-H. Ahn, “Development of the Emergency Vehicle Preemption Control System Based on UTIS,” The Journal of The Korea Institute of Intelligent Transport Systems, vol. 11, no. 2, pp. 39-47, 2012

    work page 2012

  11. [11]

    Impact of emergency vehicle preemption on signalized corridor operation: An evaluation,

    E. J. Nelson, and D. Bullock, “Impact of emergency vehicle preemption on signalized corridor operation: An evaluation,” Transportation research record, vol. 1727, no. 1, pp. 1-11, 2000

    work page 2000

  12. [12]

    Simulation testing of adaptive control, bus priority and emergency vehicle preemption in New York City

    H. Ten, Y. Qi, J. C. Falcocchio, K. -B. Kim, R. Patel, and E. Athanailos, "Simulation testing of adaptive control, bus priority and emergency vehicle preemption in New York City."

    work page

  13. [13]

    Preemption System - Emergency Vehicle - Transit,

    Eliminator. "Preemption System - Emergency Vehicle - Transit," July 27, 2021; https://collisioncontrol.net/PreemptionSystems/?gclid=Cj0KCQiApt_xBRDxARIsAAMUMu_IjJEiTe4k0tnTcFLBTB6 GNLLJ1F9WstFXn0ty2tgO5SPZwVc1syoaAqMyEALw_wcB

    work page 2021

  14. [14]

    Central Management Software (CMS) for Traffic Signal Priority Control,

    G. T. Technologies. "Central Management Software (CMS) for Traffic Signal Priority Control," July 27, 2021; www.gtt.com/project/central-management-software-cms-for-traffic-signal-priority-control/

    work page 2021

  15. [15]

    EMTRAC Systems Signal Priority: Emergency Vehicle Preemption (EVP) Utilizing ITS Functionality,

    Emtrac. "EMTRAC Systems Signal Priority: Emergency Vehicle Preemption (EVP) Utilizing ITS Functionality," July 27, 2021; https://www.emtracsystems.com/first-response/emergency-vehicle-preemption-evp/

    work page 2021

  16. [16]

    Emergency Vehicle Preemption and Priority System: Applied Information.,

    A. Information. "Emergency Vehicle Preemption and Priority System: Applied Information.," July 27, 2021; appinfoinc.com/solutions/preemption-priority/

    work page 2021

  17. [17]

    S. R. A. Guin, D. Kwesiga, G. N. S. Kumar, M. Hunter, Strategy Analysis and Evaluation for Emergency Vehicle Preemption and Transit Signal Priority with Connected Vehicles using Software in the Loop Simulation , 2023

    work page 2023

  18. [18]

    Manual on Uniform Traffic Devices for Streets and Highways (MUTCD), 11th Edition,

    "Manual on Uniform Traffic Devices for Streets and Highways (MUTCD), 11th Edition," August, 2023; https://mutcd.fhwa.dot.gov/pdfs/11th_Edition/mutcd11thedition.pdf

    work page 2023

  19. [19]

    A strategy for emergency vehicle preemption and route selection,

    K. Shaaban, M. A. Khan, R. Hamila, and M. Ghanim, “A strategy for emergency vehicle preemption and route selection,” Arabian Journal for Science and Engineering, vol. 44, no. 10, pp. 8905-8913, 2019

    work page 2019

  20. [20]

    Traffic signal priority & preemption,

    T. WSDOT. "Traffic signal priority & preemption," July 29, 2021; https://tsmowa.org/category/signal-operations/traffic- signal-priority-preemption

    work page 2021

  21. [21]

    Traffic Signal Timing Manual: Chapter 9,

    USDOT. "Traffic Signal Timing Manual: Chapter 9," July 23, 2021; https://ops.fhwa.dot.gov/publications/fhwahop08024/chapter9.htm#9.1

    work page 2021

  22. [22]

    Development of Emergency Vehicle Preemption Strategies on Smart Corridors in a Digital Twin Environment

    S. Roy, M. Hunter, A. Saroj, and A. Guin, "Development of Emergency Vehicle Preemption Strategies on Smart Corridors in a Digital Twin Environment." pp. 12-21. 16

    work page

  23. [23]

    Training ai for self -driving vehicles: the challenge of scale,

    A. Grzywaczewski, “Training ai for self -driving vehicles: the challenge of scale,” Available from Internet: https://devblogs. nvidia. com/training-self-driving-vehicles-challenge-scale, 2017

    work page 2017

  24. [24]

    Deep learning for self-driving cars: Chances and challenges

    Q. Rao, and J. Frtunikj, "Deep learning for self-driving cars: Chances and challenges." pp. 35-38

    work page

  25. [25]

    Next-generation experimentation with self-driving laboratories,

    F. Häse, L. M. Roch, and A. Aspuru-Guzik, “Next-generation experimentation with self-driving laboratories,” Trends in Chemistry, vol. 1, no. 3, pp. 282-291, 2019

    work page 2019

  26. [26]

    CAIAS simulator: self -driving vehicle simulator for AI research

    S. Hossain, A. R. Fayjie, O. Doukhi, and D. -j. Lee, "CAIAS simulator: self -driving vehicle simulator for AI research." pp. 187-195

    work page

  27. [27]

    End-to-end learning for lane keeping of self-driving cars

    Z. Chen, and X. Huang, "End-to-end learning for lane keeping of self-driving cars." pp. 1856-1860

    work page

  28. [28]

    Machine learning in transportation engineering: A feasibility study,

    T. Arciszewski, S. Khasnabis, S. K. Hoda, and W. Ziarko, “Machine learning in transportation engineering: A feasibility study,” Applied Artificial Intelligence an International Journal, vol. 8, no. 1, pp. 109-124, 1994

    work page 1994

  29. [29]

    Real -time Emergency Vehicle Event Detection Using Audio Data,

    Z. Islam, and M. Abdel -Aty, “Real -time Emergency Vehicle Event Detection Using Audio Data,” arXiv preprint arXiv:2202.01367, 2022

    work page arXiv 2022

  30. [30]

    An IoT -based intelligent traffic engagement system with emergency vehicles pre-emption,

    B. Pant, H. Sharma, R. Chawla, and C. Kumar, “An IoT -based intelligent traffic engagement system with emergency vehicles pre-emption,” International Journal of Sensor Networks, vol. 40, no. 1, pp. 10-19, 2022

    work page 2022

  31. [31]

    SMART Emergency Vehicle Management at Signalized Intersection using Machine Learning,

    R. Savithramma, R. Sumathi, and H. Sudhira, “SMART Emergency Vehicle Management at Signalized Intersection using Machine Learning,” Indian Journal of Science and Technology, vol. 15, no. 35, pp. 1754-1763, 2022

    work page 2022

  32. [32]

    On -demand greenwave for emergency vehicles in a time - varying road network with uncertainties,

    W. Min, L. Yu, P. Chen, M. Zhang, Y. Liu, and J. Wang, “On -demand greenwave for emergency vehicles in a time - varying road network with uncertainties,” IEEE Transactions on Intelligent Transportation Systems, vol. 21, no. 7, pp. 3056-3068, 2019

    work page 2019

  33. [33]

    A Survey on Emergency Vehicle Preemption Methods Based on Routing and Scheduling

    S. J. Kamble, and M. R. Kounte, “A Survey on Emergency Vehicle Preemption Methods Based on Routing and Scheduling.”

    work page

  34. [34]

    PTV Vissim 2021 user manual,

    A. PTV, “PTV Vissim 2021 user manual,” PTV AG: Karlsruhe, Germany, 2021

    work page 2021

  35. [35]

    Google Maps: Peachtree Industrial Boulevard,

    Google. "Google Maps: Peachtree Industrial Boulevard," July 27, 2021; https://www.google.com/maps/dir/33.9459033,- 84.2363882/34.0064816,-84.1701493/@33.9949555,-84.2917807,11679m/data=!3m1!1e3!4m2!4m1!3e0

    work page 2021

  36. [36]

    Learning classification models with soft-label information,

    Q. Nguyen, H. Valizadegan, and M. Hauskrecht, “Learning classification models with soft-label information,” Journal of the American Medical Informatics Association, vol. 21, no. 3, pp. 501-508, 2014

    work page 2014

  37. [37]

    Optimally splitting cases for training and testing high dimensional classifiers,

    K. K. Dobbin, and R. M. Simon, “Optimally splitting cases for training and testing high dimensional classifiers,” BMC medical genomics, vol. 4, no. 1, pp. 31, 2011

    work page 2011

  38. [38]

    5 Reasons why you should use Cross -Validation in your Data Science Projects,

    D. Shulga. "5 Reasons why you should use Cross -Validation in your Data Science Projects," March 16, 2023; https://towardsdatascience.com/5-reasons-why-you-should-use-cross-validation-in-your-data-science-project- 8163311a1e79

    work page 2023

  39. [39]

    Hold -out vs. Cross -validation in Machine Learning,

    E. Allibhai. "Hold -out vs. Cross -validation in Machine Learning," November 18, 2022; https://medium.com/@eijaz/holdout-vs-cross-validation-in-machine-learning- 7637112d3f8f#:~:text=Cross%2Dvalidation%20is%20usually%20the,just%20one%20train%2Dtest%20split

    work page 2022

  40. [40]

    Github: Peachtree -industrial-Boulevard-ML-model-Development,

    S. Roy. "Github: Peachtree -industrial-Boulevard-ML-model-Development," August 1, 2024; https://github.com/hunter- guin-gatech/Peachtree-industrial-Boulevard-ML-model-Development

    work page 2024

  41. [41]

    Emergency Vehicle Preemption Strategies using Machine Learning to Optimize Traffic Operations,

    S. Roy, “Emergency Vehicle Preemption Strategies using Machine Learning to Optimize Traffic Operations,” Dissertation, School of Civil and Environmental Engineering, Georgia Institute of Technology, 2023

    work page 2023