Pith. sign in

REVIEW 3 major objections 2 minor 66 references

Urban Deceleration Behavior Modes Under Scene Context: An Early-Kinematic Classifier from Argoverse 2 Multi-Agent Trajectories

T0 review · 3 major / 2 minor · reviewed 2026-07-02 · grok-4.3

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read Urban deceleration events cluster into four stable kinematic modes identifiable from the first second of motion.

desk verdict Four-mode taxonomy from K-means on Argoverse 2 deceleration events is new empirical work but rests on untested feature and event choices. read the letter →

arxiv 2607.00027 v1 pith:XRMRDWB6 submitted 2026-06-21 cs.RO cs.LGcs.SYeess.SPeess.SY

Eni Solomon Laughter This is my paper · ORCID
classification cs.ROcs.LGcs.SYeess.SPeess.SY
keywords urbandecelerationbehaviormodesK-meansclusteringArgoverse2kinematicclassifiercarfollowingscenecontextearlyprediction
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

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

The reading

The paper extracts over a thousand sustained deceleration events from urban driving logs in the Argoverse 2 dataset. It encodes each as a 19-dimensional vector of kinematic features and applies K-means clustering to reveal four modes that prove stable under bootstrap resampling. An early classifier predicts which mode an event belongs to using only the initial 1.0 seconds of data, achieving a macro-F1 score of 0.758, with scene context providing a modest boost. Pair age emerges as the primary scene variable influencing mode distribution, while most other context factors show little effect.

What carries the argument

K-means clustering applied to 19-dimensional kinematic feature vectors extracted from sustained deceleration events, validated by bootstrap stability analysis, and used to train an early-event HistGradientBoosting classifier.

What would settle it

Re-clustering the same 1,219 events using a different algorithm such as Gaussian mixture models or altering the feature vector to exclude jerk components, then finding low agreement with the original four-mode partition, would falsify the stability of the reported modes.

Watch

Extended reading notes

Core claim

Four stable behavioral modes structure urban deceleration: anticipatory soft braking at 62.8 percent, reactive closing at 30.6 percent, brake-like jerk at 4.8 percent, and an outlier category at 1.8 percent. These modes are recovered via K-means on 19 kinematic features with a bootstrap Adjusted Rand Index of 0.897. The modes remain consistent in medium-speed driving but shift in low-speed regimes. An early kinematic classifier reaches macro-F1 of 0.758 at one second, and pair age is the dominant contextual modulator.

Load-bearing premise

That applying K-means to the selected 19-dimensional kinematic features on the extracted events will identify distinct and interpretable behavioral modes instead of artifacts of the feature design or clustering choices.

Editorial extensions

If this is right

  • Mode membership can be predicted reliably from the first second of an event's kinematics.
  • The four modes hold across medium-speed driving conditions but vary at low speeds.
  • Scene context adds only a small improvement to classification accuracy beyond kinematics.
  • Pair age shows a medium effect on mode distribution while geometry and vulnerable road user proximity show negligible effects.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • These modes could serve as building blocks for more human-like car-following controllers in autonomous vehicles.
  • Low-speed dependence suggests the need for speed-regime-specific models in dense traffic.
  • Early detection of mode could enable proactive safety interventions in shared road scenarios.
  • The dominance of early jerk features points to jerk as a key signal for intent inference.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit.

Referee Report

3 major / 2 minor

Summary. The paper extracts 1,219 sustained deceleration events from 234 Argoverse 2 urban driving logs, encodes them as 19-dimensional kinematic feature vectors, applies K-means clustering with bootstrap stability analysis to identify four modes (anticipatory soft 62.8%, reactive closing 30.6%, brake-like jerk 4.8%, outlier 1.8%) with ARI 0.897, quantifies scene-context modulation (pair age epsilon^2=0.085 dominant), and trains a HistGradientBoosting early-event classifier achieving macro-F1=0.758 at 1.0 s (scene context adds +0.059 F1). Modes are reported as regime-invariant at medium speeds (ARI 0.817) but dependent at low speeds (ARI 0.166).

Significance. If the recovered modes are shown to be robust to upstream choices rather than artifacts, the work supplies a data-driven taxonomy for urban deceleration with direct utility for AV motion planning and car-following models. Credit is due for leveraging a public multi-agent dataset, reporting bootstrap ARI across 50 resamples, and providing an early-window classifier that isolates jerk as the dominant predictive signal.

major comments (3)
  1. [Methods (event extraction and clustering)] The central claim of four stable, interpretable modes with bootstrap ARI 0.897 rests on K-means applied to a fixed 19-dimensional kinematic feature set extracted from 1,219 events; however, the manuscript provides no justification for K-means over alternatives (GMM, hierarchical) and reports no ablation on the feature definitions, the precise criteria for 'sustained deceleration events,' or the number of clusters. This leaves the partition vulnerable to being an artifact of feature engineering or event selection rather than distinct behavioral regimes.
  2. [Results (regime dependence)] The regime analysis reports ARI=0.166 at low speed versus 0.817 at medium speed; this sensitivity directly undermines the headline assertion of 'stable modes' and 'regime-invariant' structure in medium-speed driving without additional controls or re-clustering experiments that isolate speed as a covariate.
  3. [Results (early-event classifier)] The classifier performance (macro-F1=0.758 at 1.0 s) and scene-context contribution (+0.059 F1) are presented without reported sensitivity to the kinematic feature definitions or to the clustering-derived labels themselves; because the labels are the output of the unablated K-means step, downstream predictive claims inherit the same upstream uncertainty.
minor comments (2)
  1. [Abstract/Methods] The abstract and methods should explicitly list or tabulate the 19 kinematic features and the numerical thresholds used to define 'sustained deceleration events' to support reproducibility.
  2. [Methods/Figures] Figure captions and the bootstrap procedure description would benefit from stating the exact K-means hyperparameters (initialization, distance metric, max iterations) used for the reported ARI values.

Simulated Author's Rebuttal

3 responses · 0 unresolved

We thank the referee for the constructive critique, which identifies key areas where our methodological choices require stronger justification. We address each major comment below with clarifications drawn from the manuscript and indicate revisions where the points are well-taken.

read point-by-point responses

  1. Referee: [Methods (event extraction and clustering)] The central claim of four stable, interpretable modes with bootstrap ARI 0.897 rests on K-means applied to a fixed 19-dimensional kinematic feature set extracted from 1,219 events; however, the manuscript provides no justification for K-means over alternatives (GMM, hierarchical) and reports no ablation on the feature definitions, the precise criteria for 'sustained deceleration events,' or the number of clusters. This leaves the partition vulnerable to being an artifact of feature engineering or event selection rather than distinct behavioral regimes.

    Authors: K-means was selected for its scalability on 1219 samples in 19D space and for producing directly interpretable centroids that align with kinematic intuition (e.g., jerk magnitude). The bootstrap ARI of 0.897 across 50 resamples supplies quantitative evidence against pure artifact. We nevertheless agree that explicit rationale and sensitivity checks are absent. In revision we will (i) state the rationale and cite comparable trajectory-clustering studies, (ii) report the elbow/silhouette analysis used to select k=4, and (iii) add a limitations paragraph on feature and event-definition choices. Full ablations on every upstream decision exceed the scope of the current study but will be flagged as future work. revision: partial

  2. Referee: [Results (regime dependence)] The regime analysis reports ARI=0.166 at low speed versus 0.817 at medium speed; this sensitivity directly undermines the headline assertion of 'stable modes' and 'regime-invariant' structure in medium-speed driving without additional controls or re-clustering experiments that isolate speed as a covariate.

    Authors: The manuscript already reports both ARI values and explicitly qualifies the claim: modes are regime-invariant only at medium speeds. The low-speed dependence is presented as a substantive finding rather than a flaw. To avoid any overstatement we will revise the abstract, results, and discussion to foreground the speed qualification and remove any unqualified use of 'regime-invariant.' Additional controlled re-clustering experiments that hold speed fixed are not performed in the current work; we will note this as a limitation and direction for follow-up. revision: yes

  3. Referee: [Results (early-event classifier)] The classifier performance (macro-F1=0.758 at 1.0 s) and scene-context contribution (+0.059 F1) are presented without reported sensitivity to the kinematic feature definitions or to the clustering-derived labels themselves; because the labels are the output of the unablated K-means step, downstream predictive claims inherit the same upstream uncertainty.

    Authors: We acknowledge that the classifier inherits label uncertainty from the upstream K-means step. The bootstrap stability metric offers partial mitigation by showing label consistency under resampling, yet we agree that direct sensitivity of the F1 score to feature definitions or alternative labelings is unreported. In revision we will add a short limitations subsection stating this dependency and noting that the reported scene-context gain is conditional on the current pipeline. No new classifier ablations will be added at this stage. revision: partial

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity; standard analysis of external dataset

full rationale

The derivation chain consists of event extraction from the public Argoverse 2 dataset, 19D kinematic feature encoding, K-means clustering with bootstrap ARI stability, and a downstream HistGradientBoosting classifier trained on full-event labels to predict from the first 1.0 s. No equations reduce the reported modes, ARI values, or F1 scores to quantities defined by internal fitted parameters or self-referential definitions. No self-citations, uniqueness theorems, or ansatzes are invoked. The pipeline is self-contained against external benchmarks and uses off-the-shelf algorithms without load-bearing self-referential steps.

Assumptions & free parameters 2 free parameters · 2 assumptions · 0 invented entities

The paper relies on standard unsupervised learning assumptions and a public dataset rather than introducing new physical entities or derivations. Free parameters are limited to clustering choices and feature selection; no invented entities appear.

free parameters (2)
  • number of clusters = 4
    Determined via K-means with bootstrap stability analysis on the 19-dimensional features
  • kinematic feature set
    19-dimensional vector chosen to encode deceleration events
assumptions (2)
  • domain assumption K-means clustering on kinematic trajectory features recovers distinct behavioral modes
    Core premise for mode discovery step
  • domain assumption The 1219 extracted events from Argoverse 2 are representative of urban deceleration behavior
    Underlies generalization claims

how reviews work

0 comments
Cite this review

Pith. "Pith review of Urban Deceleration Behavior Modes Under Scene Context: An Early-Kinematic Classifier from Argoverse 2 Multi-Agent Trajectories." pith.science (2026). https://pith.science/paper/XRMRDWB6

@misc{pith2026260700027,
  author       = {Pith},
  title        = {Pith review of: Urban Deceleration Behavior Modes Under Scene Context: An Early-Kinematic Classifier from Argoverse 2 Multi-Agent Trajectories},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/XRMRDWB6}},
  note         = {Machine review of arXiv:2607.00027}
}
read the original abstract

Urban deceleration is one of the most empirically studied yet least taxonomically organized behaviors in car-following research. Recent perception-equipped autonomous-vehicle datasets enable trajectory-anchored mode discovery. We extract 1,219 sustained deceleration events from 234 urban driving logs of the Argoverse 2 Sensor dataset, encode each event in a 19-dimensional kinematic feature vector, discover behavioral modes via K-means clustering with bootstrap stability analysis, and quantify modulation by eleven scene-context variables. A HistGradientBoosting classifier predicts mode membership from the first 1.0 s of each event. Four stable modes emerge with a bootstrap Adjusted Rand Index of 0.897 across 50 resamples: anticipatory soft (62.8%), reactive closing (30.6%), brake-like jerk (4.8%), and an outlier category (1.8%). Only pair age shows a medium effect (epsilon^2 = 0.085); scene geometry and vulnerable-road-user proximity show negligible effects. The early-event classifier achieves macro-F1 = 0.758 at 1.0 s, with scene context contributing +0.059 F1 over kinematics alone. Modes are regime-invariant in medium-speed driving (ARI = 0.817) but regime-dependent at low speed (ARI = 0.166). A small set of stable kinematic modes structures urban deceleration; early-window jerk dominates predictive signal; and pair age is the primary contextual modulator.

Figures

Figures reproduced from arXiv: 2607.00027 by the authors.

Figure 1
Figure 1. Illustrative view of the Argoverse 2 Sensor dataset annotation structure. Left: bird's￾eye-view BEV LiDAR point cloud with ground-truth 3D cuboid annotations for vehicles, pedestri￾ans, and other agents (green/red boxes). Right: synchronized front-facing camera frame at an urban intersection with class-labelled bounding-box overlays. The cuboid annotations provide the per￾frame multi-agent trajectory ground truth us… view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

66 extracted references · 66 canonical work pages

  1. [1]

    Effect of Situation Kinematics on Drivers’ Rear -End Collision Avoidance Behaviour-A Combined Effect of Visual Looming, Speed, and Distance Analysis

    Xue, Q.; Ouyang, X.; Zhao, Y.; Guo, W. Effect of Situation Kinematics on Drivers’ Rear -End Collision Avoidance Behaviour-A Combined Effect of Visual Looming, Speed, and Distance Analysis. SUSTAINABILITY 2022, 14, doi:10.3390/su142215103

    work page doi:10.3390/su142215103 2022

  2. [2]

    Predicting Driver Reaction Time and Deceleration: Comparison of Perception-Reac- tion Thresholds and Evidence Accumulation Framework

    Durrani, U.; Lee, C.; Shah, D. Predicting Driver Reaction Time and Deceleration: Comparison of Perception-Reac- tion Thresholds and Evidence Accumulation Framework. ACCIDENT ANALYSIS AND PREVENTION 2021, 149, doi:10.1016/j.aap.2020.105889

    work page doi:10.1016/j.aap.2020.105889 2021

  3. [3]

    Detection and Response to Critical Lead Vehicle Deceleration Events with Pe- ripheral Vision: Glance Response Times Are Independent of Visual Eccentricity

    Svärd, M.; Bärgman, J.; Victor, T. Detection and Response to Critical Lead Vehicle Deceleration Events with Pe- ripheral Vision: Glance Response Times Are Independent of Visual Eccentricity. ACCIDENT ANALYSIS AND PREVENTION 2021, 150, doi:10.1016/j.aap.2020.105853

    work page doi:10.1016/j.aap.2020.105853 2021

  4. [4]

    Argoverse 2: Next Generation Datasets for Self-Driving Perception and Forecasting 2023

    Wilson, B.; Qi, W.; Agarwal, T.; Lambert, J.; Singh, J.; Khandelwal, S.; Pan, B.; Kumar, R.; Hartnett, A.; Pontes, J.K.; et al. Argoverse 2: Next Generation Datasets for Self-Driving Perception and Forecasting 2023

    work page 2023

  5. [5]

    An Improved Car-Following Model Considering the Influence of Space Gap to the Response

    Li, X.; Luo, X.; He, M.; Chen, S. An Improved Car-Following Model Considering the Influence of Space Gap to the Response. PHYSICA A -STATISTICAL MECHANICS AND ITS APPLICATIONS 2018, 509, 536 –545, doi:10.1016/j.physa.2018.06.069

    work page doi:10.1016/j.physa.2018.06.069 2018

  6. [6]

    Weather and Road Geometry Impact on Longitudinal Driving Behavior: Ex- ploratory Analysis Using an Empirically Supported Acceleration Modeling Framework

    Hamdar, S.; Qin, L.; Talebpour, A. Weather and Road Geometry Impact on Longitudinal Driving Behavior: Ex- ploratory Analysis Using an Empirically Supported Acceleration Modeling Framework. TRANSPORTATION RE- SEARCH PART C-EMERGING TECHNOLOGIES 2016, 67, 193–213, doi:10.1016/j.trc.2016.01.017

    work page doi:10.1016/j.trc.2016.01.017 2016

  7. [7]

    A Large - Scale Dataset of Distributed Renewable Energy Scenarios on the IEEE - 33 Bus Network,

    Chen, X.; Zhu, M.; Chen, K.; Wang, P.; Lu, H.; Zhong, H.; Han, X.; Wang, X.; Wang, Y. FollowNet: A Comprehen- sive Benchmark for Car -Following Behavior Modeling. SCIENTIFIC DATA 2023, 10, doi:10.1038/s41597 -023- 02718-7

    work page doi:10.1038/s41597 2023

  8. [8]

    A Recurrent Neural Network Model for Predicting Two -Leader Car -Following Behavior

    Das, S.; Maurya, A.K.; Dey, A. A Recurrent Neural Network Model for Predicting Two -Leader Car -Following Behavior. TRANSPORTATION LETTERS -THE INTERNATIONAL JOURNAL OF TRANSPORTATION RE- SEARCH 2024, 16, 461–475, doi:10.1080/19427867.2023.2205190

    work page doi:10.1080/19427867.2023.2205190 2024

Show all 66 references
  1. [9]

    A Driving -Preference-Aware Framework for Vehicle Lane Change Prediction

    Lyu, Y.; Wang, Y.; Liu, H.; Dong, X.; He, Y.; Ren, Y. A Driving -Preference-Aware Framework for Vehicle Lane Change Prediction. SENSORS 2025, 25, doi:10.3390/s25175342

    2025 doi

  2. [10]

    A Review of Car -Following and Lane -Changing Models under Heterogeneous Environments

    Chen, Y.; Dong, C.; Lyu, K.; Shi, X.; Han, G.; Wang, H. A Review of Car -Following and Lane -Changing Models under Heterogeneous Environments. PHYSICA A-STATISTICAL MECHANICS AND ITS APPLICATIONS 2024, 654, doi:10.1016/j.physa.2024.130127

    2024 doi

  3. [11]

    Analyzing and Modeling Drivers’ Deceleration Behavior from Normal Driving

    Deligianni, S.P.; Quddus, M.; Morris, A.; Anvuur, A.; Reed, S. Analyzing and Modeling Drivers’ Deceleration Behavior from Normal Driving. TRANSPORTATION RESEARCH RECORD 2017, 134–141

    2017

  4. [12]

    Learning the Driver Acceleration/Deceleration Behavior Under High -Speed Environments From Naturalistic Driving Data

    Liu, C.; Zhang, W. Learning the Driver Acceleration/Deceleration Behavior Under High -Speed Environments From Naturalistic Driving Data. IEEE INTELLIGENT TRANSPORTATION SYSTEMS MAGAZINE 2022, 14, 78–91, doi:10.1109/MITS.2020.3014115

    2022 doi

  5. [13]

    Exploring Transition Durations of Rear-End Collisions Based on Vehicle Trajectory Data: A Survival Modeling Approach

    Li, Y.; Wu, D.; Chen, Q.; Lee, J.; Long, K. Exploring Transition Durations of Rear-End Collisions Based on Vehicle Trajectory Data: A Survival Modeling Approach. ACCIDENT ANALYSIS AND PREVENTION 2021, 159, doi:10.1016/j.aap.2021.106271. 21 of 23

    2021 doi

  6. [14]

    Biologically Guided Driver Modeling: The Stop Behavior of Hu- man Car Drivers

    Da Lio, M.; Mazzalai, A.; Gurney, K.; Saroldi, A. Biologically Guided Driver Modeling: The Stop Behavior of Hu- man Car Drivers. IEEE TRANSACTIONS ON INTELLIGENT TRANSPORTATION SYSTEMS 2018, 19, 2454–2469, doi:10.1109/TITS.2017.2751526

    2018 doi

  7. [15]

    Analysis of Drivers’ Deceleration Behavior Based on Naturalistic Driving Data

    Li, S.; Li, P.; Yao, Y.; Han, X.; Xu, Y.; Chen, L. Analysis of Drivers’ Deceleration Behavior Based on Naturalistic Driving Data. TRAFFIC INJURY PREVENTION 2020, 21, 42–47

    2020

  8. [16]

    Effect of Daily Car -Following Behaviors on Urban Roadway Rear - End Crashes and near -Crashes: A Naturalistic Driving Study

    Wang, X.; Zhang, X.; Guo, F.; Gu, Y.; Zhu, X. Effect of Daily Car -Following Behaviors on Urban Roadway Rear - End Crashes and near -Crashes: A Naturalistic Driving Study. ACCIDENT ANALYSIS AND PREVENTION 2022, 164

    2022

  9. [17]

    Car-Following Behavior Based on LiDAR Trajectory Data at Urban Intersections

    Yang, Z.; Zheng, J.; Yu, W.; Xiao, Y.; Li, J.; Li, T. Car-Following Behavior Based on LiDAR Trajectory Data at Urban Intersections. IEEE SYSTEMS JOURNAL 2024, 18, 438–449

    2024

  10. [18]

    Classifying Rapid Deceleration Events for Older Drivers

    Yamagishi, M. Classifying Rapid Deceleration Events for Older Drivers. PROCEEDINGS OF THE 22ND CON- GRESS OF THE INTERNATIONAL ERGONOMICS ASSOCIATION, VOL 4, IEA 2024 2025, 56, 215–221

    2024

  11. [19]

    Driver Influence on Vehicle Trajectory Prediction

    Khakzar, M.; Bond, A.; Rakotonirainy, A.; Oviedo -Trespalacios, O.; Dehkordi, S.G. Driver Influence on Vehicle Trajectory Prediction. ACCIDENT ANALYSIS AND PREVENTION 2021, 157

    2021

  12. [20]

    Chen, K.-T.; Chen, H. -Y.W. Driving Style Clustering Using Naturalistic Driving Data. TRANSPORTATION RE- SEARCH RECORD 2019, 2673, 176–188

    2019

  13. [21]

    Clustering of Driving Encounter Scenarios Using Connected Vehicle Trajectories

    Wang, W.; Ramesh, A.; Zhu, J.; Li, J.; Zhao, D. Clustering of Driving Encounter Scenarios Using Connected Vehicle Trajectories. IEEE TRANSACTIONS ON IN℡LIGENT VEHICLES 2020, 5, 485–496

    2020

  14. [22]

    Dynamic Clustering Analysis for Driving Styles Identi- fication

    de Zepeda, M.V.N.; Meng, F.; Su, J.; Zeng, X.-J.; Wang, Q. Dynamic Clustering Analysis for Driving Styles Identi- fication. ENGINEERING APPLICATIONS OF ARTIFICIAL IN℡LIGENCE 2021, 97

    2021

  15. [23]

    Normal and Risky Driving Patterns Identification in Clear and Rainy Weather on Freeway Segments Using Vehicle Kinematics Trajectories and Time Series Cluster Analysis

    Ali, E.; Ahmed, M.; Yang, G. Normal and Risky Driving Patterns Identification in Clear and Rainy Weather on Freeway Segments Using Vehicle Kinematics Trajectories and Time Series Cluster Analysis. IATSS RESEARCH 2021, 45, 137–152, doi:10.1016/j.iatssr.2020.07.002

    2021 doi

  16. [24]

    Driving Behavior Classification at Signalized Intersections Using Vehicle Kinematics: Application of Unsupervised Machine Learning

    Khanfar, N.O.; Elhenawy, M.; Ashqar, H., I.; Hussain, Q.; Alhajyaseen, W.K.M. Driving Behavior Classification at Signalized Intersections Using Vehicle Kinematics: Application of Unsupervised Machine Learning. INTERNA- TIONAL JOURNAL OF INJURY CONTROL AND SAFETY PROMOTION 2023...

    2023

  17. [25]

    Characterizing Driver Behavior Using Naturalistic Driving Data

    Lee, J.; Jang, K. Characterizing Driver Behavior Using Naturalistic Driving Data. ACCIDENT ANALYSIS AND PREVENTION 2024, 208, doi:10.1016/j.aap.2024.107779

    2024 doi

  18. [26]

    A High -Resolution Trajectory Data Driven Approach for Vehicle Conflict Risk Prediction Considering Driving Style

    Pan, Y.; Zhu, D.; Ding, W.; Song, N.; Zhang, W. A High -Resolution Trajectory Data Driven Approach for Vehicle Conflict Risk Prediction Considering Driving Style. PROCEEDINGS OF THE INSTITUTION OF MECHANICAL ENGINEERS PART D-JOURNAL OF AUTOMOBILE ENGINEERING 2025

    2025

  19. [27]

    Identification and Interpretation of Aggressive Driving Behavior under Time Pressure

    Cui, T.; Ma, Y.; Chen, S.; Xing, G.; Zhang, H.; Ong, G.P. Identification and Interpretation of Aggressive Driving Behavior under Time Pressure. TRANSPORTATION RESEARCH RECORD 2026, 2680, 406–420

    2026

  20. [28]

    Driving Behavior Risk Measurement and Cluster Analysis Driven by Vehicle Trajectory Data

    Chen, S.; Cheng, K.; Yang, J.; Zang, X.; Luo, Q.; Li, J. Driving Behavior Risk Measurement and Cluster Analysis Driven by Vehicle Trajectory Data. APPLIED SCIENCES-BASEL 2023, 13

    2023

  21. [29]

    The Characteristics of Driver Lane -Changing Behaviour in Congested Road Environments

    Wang, W.; Cheng, G. The Characteristics of Driver Lane -Changing Behaviour in Congested Road Environments. TRANSPORTATION SAFETY AND ENVIRONMENT 2024, 6

    2024

  22. [30]

    Quantifying Visual Road Environment to Establish a Speeding Prediction Model: An Examination Using Naturalistic Driving Data

    Yu, B.; Chen, Y.; Bao, S. Quantifying Visual Road Environment to Establish a Speeding Prediction Model: An Examination Using Naturalistic Driving Data. Accident Analysis & Prevention 2019, 129, 289 –298, doi:10.1016/j.aap.2019.05.011

    2019 doi

  23. [31]

    Guo, H.; Xie, K.; Keyvan-Ekbatani, M.; IEEE Lane Change Detection Using Naturalistic Driving Data.; 2021

    2021

  24. [32]

    Vehicle Lane Change Decision Model Based on Random Forest

    Gu, X.; Yu, J.; Han, Y.; Han, M.; Wei, L. Vehicle Lane Change Decision Model Based on Random Forest. In Pro- ceedings of the 2019 IEEE International Conference on Power, Intelligent Computing and Systems (ICPICS); July 2019; pp. 115–120. 22 of 23

    2019

  25. [33]

    Exploratory Analysis of Lane Changing on Freeways Based on Driver Be- havior

    Hill, C.; Elefteriadou, L.; Kondyli, A. Exploratory Analysis of Lane Changing on Freeways Based on Driver Be- havior. JOURNAL OF TRANSPORTATION ENGINEERING 2015, 141

    2015

  26. [34]

    Heterogeneity in Car -Following Behavior: Theory and Empirics

    Ossen, S.; Hoogendoorn, S.P. Heterogeneity in Car -Following Behavior: Theory and Empirics. TRANSPORTA- TION RESEARCH PART C-EMERGING TECHNOLOGIES 2011, 19, 182–195

    2011

  27. [35]

    Driver Behavior Classification: A Systematic Literature Review

    Bouhsissin, S.; Sael, N.; Benabbou, F. Driver Behavior Classification: A Systematic Literature Review. IEEE Access 2023, 11, 14128–14153, doi:10.1109/ACCESS.2023.3243865

    2023 doi

  28. [36]

    Human-like Car-Following Model for Autonomous Vehicles Considering the Cut- in Behavior of Other Vehicles in Mixed Traffic

    Fu, R.; Li, Z.; Sun, Q.; Wang, C. Human-like Car-Following Model for Autonomous Vehicles Considering the Cut- in Behavior of Other Vehicles in Mixed Traffic. ACCIDENT ANALYSIS AND PREVENTION 2019, 132

    2019

  29. [37]

    -W.; Vega, J.H.; Tsiang, N.R

    Hu, J.; Lin, B.T. -W.; Vega, J.H.; Tsiang, N.R. -L. Predictive Models of Driver Deceleration and Acceleration Re- sponses to Lead Vehicle Cutting In and Out. TRANSPORTATION RESEARCH RECORD 2023, 2677, 92–102

    2023

  30. [38]

    A Car-Following Model Considering Preceding Vehicle’s Lane-Chang- ing Process

    Zhao, M.; Wang, S.-H.; Sun, D.; Wang, X.-J. A Car-Following Model Considering Preceding Vehicle’s Lane-Chang- ing Process. IEEE ACCESS 2019, 7, 89913–89923

    2019

  31. [39]

    Exploring the Stimulative Effect on Following Drivers in a Consecutive Lane Change Using Microscopic Vehicle Trajectory Data

    Gu, R.; Li, Y.; Cen, X. Exploring the Stimulative Effect on Following Drivers in a Consecutive Lane Change Using Microscopic Vehicle Trajectory Data. TRANSPORTATION SAFETY AND ENVIRONMENT 2023, 5

    2023

  32. [40]

    Influence of Surrounding Traffic on Lane Change Dynamics: In- sights from a Video-Based Laboratory Study

    Jokhio, S.; Dürr, M.; Bärgman, J.; Baumann, M. Influence of Surrounding Traffic on Lane Change Dynamics: In- sights from a Video-Based Laboratory Study. TRANSPORTATION RESEARCH PART F-TRAFFIC PSYCHOLOGY AND BEHAVIOUR 2024, 105, 87–98, doi:10.1016/j.trf.2024.06.025

    2024 doi

  33. [41]

    Categorization of the Lane Change Decision Process on Freeways

    Keyvan-Ekbatani, M.; Knoop, V.; Daamen, W. Categorization of the Lane Change Decision Process on Freeways. TRANSPORTATION RESEARCH PART C -EMERGING TECHNOLOGIES 2016, 69, 515 –526, doi:10.1016/j.trc.2015.11.012

    2016 doi

  34. [42]

    Empirics and Models of Fragmented Lane Changes

    Mullakkal-Babu, F.; Wang, M.; van Arem, B.; Happee, R. Empirics and Models of Fragmented Lane Changes. IEEE OPEN JOURNAL OF INTELLIGENT TRANSPORTATION SYSTEMS 2020, 1, 187 –200, doi:10.1109/OJITS.2020.3029056

    2020 doi

  35. [43]

    Kinematic Cues in Driver-Pedestrian Communication to Support Safe Road Crossing

    Noonan, Z.; Gershon, P.; Domeyer, J.; Mehler, B.; Reimer, B. Kinematic Cues in Driver-Pedestrian Communication to Support Safe Road Crossing. ACCIDENT ANALYSIS AND PREVENTION 2023, 192

    2023

  36. [44]

    How Deceleration Onset Time and Vehicle Speed Shape Pedestrian Per- ceived Safety and Comfort in Yielding Autonomous Vehicles

    Chen, W.; Jiang, Q.; Zhuang, X.; Ma, G. How Deceleration Onset Time and Vehicle Speed Shape Pedestrian Per- ceived Safety and Comfort in Yielding Autonomous Vehicles. ENGINEERING PSYCHOLOGY AND COGNITIVE ERGONOMICS, EPCE 2025, PT I 2025, 15776, 165–174

    2025

  37. [46]

    A Feature Learning Approach Based on XGBoost for Driving Assessment and Risk Prediction

    Shi, X.; Wong, Y.; Li, M.; Palanisamy, C.; Chai, C. A Feature Learning Approach Based on XGBoost for Driving Assessment and Risk Prediction. ACCIDENT ANALYSIS AND PREVENTION 2019, 129, 170 –179, doi:10.1016/j.aap.2019.05.005

    2019 doi

  38. [47]

    Recognition and Interpretation of Aggressive Driving Behav- ior for Heavy -Duty Vehicles Based on Artificial Neural Network and SHAP

    Cheng, C.; Chen, S.; Ma, Y.; Khattak, A.J.; Zhang, Z. Recognition and Interpretation of Aggressive Driving Behav- ior for Heavy -Duty Vehicles Based on Artificial Neural Network and SHAP. HUMAN FACTORS AND ERGO- NOMICS IN MANUFACTURING & SERVICE INDUSTRIES 2024, 34, 177–189

    2024

  39. [49]

    Autonomous Vehicle Lane -Changing Dynamics and Impact on the Immediate Follower

    Ali, Y. Autonomous Vehicle Lane -Changing Dynamics and Impact on the Immediate Follower. ANALYTIC METHODS IN ACCIDENT RESEARCH 2025, 46, doi:10.1016/j.amar.2025.100388

    2025 doi

  40. [50]

    Arterial Vehicle Trajectory Re- construction by Integrating Driving Behaviour Pattern Recognition

    Ma, J.; Yang, Y.; Ren, G.; Wang, S.; Cao, Q.; Song, J.; Deng, Y.; Wang, R.; Lin, W. Arterial Vehicle Trajectory Re- construction by Integrating Driving Behaviour Pattern Recognition. TRANSPORTMETRICA B-TRANSPORT DY- NAMICS 2025, 13, doi:10.1080/21680566.2025.2575955. 23 of 23

    2025 doi

  41. [51]

    -J.; Kim, D

    Joo, Y.-J.; Kim, E. -J.; Kim, D. -K.; Park, P.Y. A Generalized Driving Risk Assessment on High -Speed Highways Using Field Theory. ANALYTIC METHODS IN ACCIDENT RESEARCH 2023, 40, doi:10.1016/j.amar.2023.100303

    2023 doi

  42. [52]

    F -F Diagram: A Two -Dimensional Surrogate Safety Indicator Based on Field Theory for Lane Change Risk Assessment

    Liu, Z.; Xiang, Q. F -F Diagram: A Two -Dimensional Surrogate Safety Indicator Based on Field Theory for Lane Change Risk Assessment. IEEE TRANSACTIONS ON IN ℡LIGENT TRANSPORTATION SYSTEMS 2025, 26, 10694–10709, doi:10.1109/TITS.2025.3547969

    2025 doi

  43. [53]

    A Longitudinal Car -Following Risk Assessment Model Based on Risk Field Theory for Autonomous Vehicles

    Wu, B.; Yan, Y.; Ni, D.; Li, L. A Longitudinal Car -Following Risk Assessment Model Based on Risk Field Theory for Autonomous Vehicles. INTERNATIONAL JOURNAL OF TRANSPORTATION SCIENCE AND TECHNOLOGY 2021, 10, 60–68, doi:10.1016/j.ijtst.2020.05.005

    2021 doi

  44. [54]

    An Improved Risk Estimation Model of Lane Change Using Naturalistic Vehicle Trajectories

    Xue, Q.; Wang, K.; Lu, J.J.; Xing, Y.; Gu, X.; Zhang, M. An Improved Risk Estimation Model of Lane Change Using Naturalistic Vehicle Trajectories. JOURNAL OF TRANSPORTATION SAFETY & SECURITY 2023, 15, 963–986

    2023

  45. [55]

    Non-Stationary Crash Risk Modelling of Powered Two-Wheelers Using Extreme Value Analysis of Surrogate Crash Events

    Kar, P.; Venthuruthiyil, S.P.; Chunchu, M. Non-Stationary Crash Risk Modelling of Powered Two-Wheelers Using Extreme Value Analysis of Surrogate Crash Events. ACCIDENT ANALYSIS AND PREVENTION 2023, 183, doi:10.1016/j.aap.2023.106973

    2023 doi

  46. [56]

    (Michael); Zhou, X

    Li, L.; Jiang, R.; He, Z.; Chen, X. (Michael); Zhou, X. Trajectory Data-Based Traffic Flow Studies: A Revisit. TRANS- PORTATION RESEARCH PART C-EMERGING TECHNOLOGIES 2020, 114, 225–240, doi:10.1016/j.trc.2020.02.016

    2020 doi

  47. [57]

    Vehicle Trajectory Dataset from Drone Videos Including Off-Ramp and Congested Traffic - Analysis of Data Quality, Traffic Flow, and Accident Risk

    Berghaus, M.; Lamberty, S.; Ehlers, J.; Kallo, E.; Oeser, M. Vehicle Trajectory Dataset from Drone Videos Including Off-Ramp and Congested Traffic - Analysis of Data Quality, Traffic Flow, and Accident Risk. COMMUNICA- TIONS IN TRANSPORTATION RESEARCH 2024, 4

    2024

  48. [58]

    Global Lessons Learned from Naturalistic Driving Studies to Advance Traffic Safety and Operation Research: A Systematic Review

    Ahmed, M.M.; Khan, M.N.; Das, A.; Dadvar, S.E. Global Lessons Learned from Naturalistic Driving Studies to Advance Traffic Safety and Operation Research: A Systematic Review. Accident Analysis & Prevention 2022, 167, 106568

    2022

  49. [59]

    Self -Supervised Transformer for Trajectory Prediction Using Noise Imputed Past Trajectory

    Bharilya, V.; Arora, A.; Kumar, N. Self -Supervised Transformer for Trajectory Prediction Using Noise Imputed Past Trajectory. IEEE TRANSACTIONS ON IN ℡LIGENT TRANSPORTATION SYSTEMS 2025, 26, 8454 –8466, doi:10.1109/TITS.2025.3550711

    2025 doi

  50. [60]

    A Systematic Review of Machine Learning -Based Micro- scopic Traffic Flow Models and Simulations

    Rowan, D.; He, H.; Hui, F.; Yasir, A.; Mohammed, Q. A Systematic Review of Machine Learning -Based Micro- scopic Traffic Flow Models and Simulations. COMMUNICATIONS IN TRANSPORTATION RESEARCH 2025, 5

    2025

  51. [61]

    Learning the Representation of Surrogate Safety Measures to Identify Traffic Conflict

    Lu, J.; Grembek, O.; Hansen, M. Learning the Representation of Surrogate Safety Measures to Identify Traffic Conflict. ACCIDENT ANALYSIS AND PREVENTION 2022, 174, doi:10.1016/j.aap.2022.106755

    2022 doi

  52. [62]

    Consistency Analysis of Drivers’ Car -Following Behaviors

    Qin, D.; Wang, X.; Tarko, A.P.P.; Guo, Q.; Lizarazo -Jimenez, C. Consistency Analysis of Drivers’ Car -Following Behaviors. JOURNAL OF TRANSPORTATION ENGINEERING PART A-SYSTEMS 2023, 149

    2023

  53. [63]

    Kinematic Discriminants of Deceleration Behavior Modes in Car-Following: Evidence from NGSIM Trajectory Data 2026

    Laughter, E.S. Kinematic Discriminants of Deceleration Behavior Modes in Car-Following: Evidence from NGSIM Trajectory Data 2026

    2026

  54. [64]

    Trust, but Verify: Cross -Modality Fusion for HD Map Change Detection

    Lambert, J.; Hays, J. Trust, but Verify: Cross -Modality Fusion for HD Map Change Detection. In Proceedings of the Proceedings of the Neural Information Processing Systems Track on Datasets and Benchmarks (NeurIPS Da- tasets and Benchmarks 2021); 2021

    2021

  55. [65]

    On the Maximum of Cramer’s V 2025

    Hamada, E. On the Maximum of Cramer’s V 2025

    2025

  56. [66]

    Sage Research Methods - Encyclopedia of Research Design - Kruskal–Wallis Test Available online: https://meth- ods.sagepub.com/ency/edvol/encyc-of-research-design/chpt/kruskal-wallis-test (accessed on 19 June 2026)

    2026

  57. [67]

    Eta Squared, Partial Eta Squared, and Misreporting of Effect Size in Communication Research

    Levine, T.R.; Hullett, C.R. Eta Squared, Partial Eta Squared, and Misreporting of Effect Size in Communication Research. Human Comm Res 2002, 28, 612–625, doi:10.1111/j.1468-2958.2002.tb00828.x

    2002 doi

  58. [68]

    Elementary Proofs of Several Results on False Discovery Rate 2025

    Wang, R. Elementary Proofs of Several Results on False Discovery Rate 2025

    2025

Pith tools

Reviewed July 2, 2026 · model on record in the stance chip above.