Pith. sign in

REVIEW 4 major objections 7 minor 28 references

Comparison of Lightweight Methods for Vehicle Dynamics-Based Driver Drowsiness Detection

T0 review · 4 major / 7 minor · reviewed 2026-08-07 · deepseek-v4-flash

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

Pith's one-line read A simple random forest model, evaluated alongside three re-implemented published methods on a public dataset, achieves the best drowsiness-detection accuracy (88%) and shows the published methods' claimed performance does not reproduce.

desk verdict A genuinely useful reproducibility benchmark for vehicle-dynamics drowsiness detection, undermined mainly by an under-specified data split that could favor the proposed RF baseline. read the letter →

arxiv 2506.07014 v1 pith:Q46F45TQ submitted 2025-06-08 cs.LG

Yutaro Nakagama , Daisuke Ishii , Kazuki Yoshizoe This is my paper · ORCID
classification cs.LG
keywords driverdrowsinessdetectionvehicledynamicsrandomforestSVMLSTMreproducibilitydataleakageMMDAPdataset
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

This paper tries to establish that the performance of established vehicle-dynamics-based driver drowsiness detection (DDD) methods is not reproducible when the methods are re-implemented and evaluated under a common, standardized protocol on the public MMDAP dataset, and that in that same protocol a simple random forest (RF) method beats all three re-implementations. Re-implementing SvmA, SvmW, and Lstm, the authors obtain accuracies of 65%, 49%, and 70% in the standard configuration, far below the 98%, 95%, and 91% reported in the original papers; the proposed RF reaches 88% accuracy and 0.85 AUC. The authors attribute the gaps to non-standard evaluation practices in the original studies, including possible data leakage and reporting of training metrics. If true, the paper provides a strong argument for standardizing DDD evaluation and shows that simple methods on standard signals deserve attention as baselines.

What carries the argument

The paper's central mechanism is a configurable DDD framework that parameterizes the machine-learning pipeline by data-split, labeling source, and hyperparameters, enabling a common evaluation. The proposed RF method combines all feature sets used by the three prior methods (36 statistical steering features, 8 wavelet energy features, and 15 summary features from five signals) and applies univariate feature selection by ANOVA F-values before training the random forest classifier; ground-truth labels come from EEG spectral ratios, with DRT events as an alternative label source.

What would settle it

Re-run the four methods on a participant-disjoint split of MMDAP (all windows from each driver assigned to only one partition). If RF's accuracy and AUC fall to the level of the other methods, the central claim that RF is best under a fair protocol would be undermined.

Watch

Extended reading notes

Core claim

On the public MMDAP dataset, the authors re-implement three published lightweight vehicle-dynamics DDD methods (SvmA, SvmW, Lstm) and propose RF, which aggregates the feature sets of all three methods with an ANOVA F-value-based feature selection, trained with EEG-derived labels. Under configuration C2—an 8:1:1 train/validation/test split, 3-second windows with 50% overlap, 60 Hz sampling, and EEG-based drowsiness labels—RF achieves the highest AUC (0.85) and accuracy (88%) among the four methods. The re-implemented prior methods attain only 65% accuracy (SvmA), 49% accuracy (SvmW), and 70% accuracy (Lstm), whereas the original papers reported 98%, 95%, and 91%, respectively. The authors conclude that the prior results are largely due to non-standard evaluation protocols—likely training accuracy or data leakage—and that RF under a standardized protocol is a competitive and reliable baseline.

Load-bearing premise

The evaluation assumes that splitting the dataset into 8:1:1 training, validation, and test sets prevents data leakage, meaning no overlapping time windows and no data from the same participant appear in both training and test, but the paper does not describe participant-level separation.

Editorial extensions

If this is right

  • The four-way comparison provides a reproducible baseline: a simple random forest on vehicle dynamics yields 88% accuracy and 0.85 AUC on MMDAP under a standardized protocol.
  • The published accuracy claims for the three re-implemented methods (98%, 95%, 91%) are not achieved under the common configuration, indicating those reported values likely stem from training accuracy or non-standard evaluation practices.
  • Using a public dataset and standard metrics allows DDD evaluation to be standardized, and the proposed framework supports multiple configurations and feature sets for future comparisons.
  • EEG-based labeling in a straight-road simulator scenario offers one fair labeling protocol, but the same framework's results may change with other label sources such as DRT events.

Reading between the lines

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

  • The paper does not specify whether the 8:1:1 split is participant-disjoint or purely per-window; a per-window random split could let the same driver's data appear in both training and test, inflating accuracy and possibly favoring the RF, so a participant-disjoint re-split is a direct test of the claim.
  • Because the RF uses strictly more features than the other methods, some of its advantage may come from feature availability rather than the classifier; testing SvmA, SvmW, and Lstm with the same full feature set would isolate that contribution.
  • The EEG labeling rule discards the middle 17.8% of the drowsiness distribution (upper 22.2% drowsy, lower 60% awake), making the remaining two classes easier to separate, so the reported 88% is specific to that labeling cutoff.
  • The released framework could serve as a benchmark harness that future DDD papers adopt to report standardized metrics, similar to fixed evaluation protocols in other applied ML fields.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

4 major / 7 minor

Summary. The paper proposes a common framework for vehicle-dynamics-based driver drowsiness detection (DDD) using the public MMDAP dataset, re-implements three published lightweight methods (SvmA, SvmW, Lstm) and a proposed random-forest method (RF), and compares them under a common configuration C2 (8:1:1 split, 3-second windows, 50% overlap, 60 Hz). The central claim is that the RF method achieves the best performance (accuracy 88%, AUC 0.85), while the three re-implementations fall far below their originally reported values, suggesting that previously reported DDD results are not reproducible and that the field needs standardized evaluation.

Significance. If the comparison is valid, the paper provides a useful open benchmark and a reproducible baseline for vehicle-dynamics-based DDD. The authors publicly release their implementation, which is a concrete strength, and they make a good-faith effort to check reproducibility of prior work on a single public dataset. The paper also honestly lists threats to validity, including possible hyperparameter bias and dataset generalizability. However, the significance is limited by the single straight-road simulator dataset, the lack of uncertainty quantification, and the unresolved split protocol; these issues affect whether the reported ranking of RF over the other methods generalizes beyond the specific experimental setup.

major comments (4)
  1. [Section IV-C2, Table V] The C2 split, which produces the results in Table V, is described only as 'we split the dataset at a 8:1:1 ratio' with no statement about whether the split is participant-level or window-level. Because C2 uses 3-second windows with 50% overlap, a window-level random split puts overlapping 1.5-second segments from the same driver into both training and test sets; vehicle dynamics and EEG-derived labels are strongly autocorrelated within a session, so this violates the independence assumption and can inflate the reported accuracy and AUC. This is a load-bearing issue because the paper's central claim is that RF is best under C2, and the paper itself criticizes prior work for data leakage (Sect. I) and lists individual differences as a threat (Sect. VI-D). Please specify how participants and overlapping windows were assigned to the three folds; if the split was not grouped by participant, rerun the comparison with a participant-level split (e.g., group all windows of a participant into one set) and report the new metrics and ranking.
  2. [Section V, Table V] The comparison reports single point estimates (RF accuracy 88%, AUC 0.85) from one split, without error bars, confidence intervals, or significance tests. The qualitative claim that RF is 'the highest' among the four methods is therefore a statement about one random realization of the 8:1:1 split. Given that the paper's purpose is to provide a reliable comparison, please report repeated random splits or cross-validation with confidence intervals, and add a paired test (e.g., McNemar's test on the test set) between RF and each baseline.
  3. [Section III-D.4, Section VI-A] The RF method description says the top-ranked features are selected with SelectKBest but does not state the number K, and the Optuna tuning step (Sect. III-E) is described without reporting search spaces or selected hyperparameters. Section VI-A explicitly concedes that hyperparameters may not have been equally optimized and that this could make the comparison unfair. Since Table V is the paper's central result, please report K and the hyperparameter configurations for all four methods (or provide a versioned pointer to the code), or temper the comparative conclusion to match the admitted uncertainty.
  4. [Section III-C] The EEG labeling rule classifies the lower 60% and upper 22.2% of 'the level distribution' as awake and drowsy, but the paper does not say whether this distribution is computed per participant, per session, or globally across the dataset. This choice changes the learning task: per-participant quantiles make the labels rank-based within a driver, while a global distribution makes the threshold depend on dataset composition. Please state the grouping used for the quantiles and, if it is per-participant, discuss the implications for generalization to new drivers.
minor comments (7)
  1. [Abstract] The abstract contains typographical errors: 'frameworkiscarefullydesigned' is missing spaces and 'onfigurations' should be 'configurations'.
  2. [Section III-D.1] The text says 'It uses an SMV model'; this should be 'SVM'.
  3. [Section IV-B.1 and Table III] Table III lists the SvmW input as θ at 25 Hz, but Sect. IV-B.1 states the sampling rate was set to 60 Hz; please clarify which rate was used for SvmW in C1 and C2.
  4. [Section V-A] The paragraph about Lstm repeats itself: 'the recall rate was relatively low' and later 'showed a recall as low as 5%'; consider merging the two sentences.
  5. [Section I] The statement that 'it was necessary to use test data to achieve a similar level of performance' for [4] is vague; specify whether this refers to the C1 training-configuration evaluation or another protocol.
  6. [Figure 4] The figure panels are labeled (a) and (b), but the text refers to 'above' and 'below'; please make the references consistent.
  7. [Section III-E] For archival reproducibility, please include a specific commit hash or versioned DOI for the public repository, in addition to the URL.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the paper's comparison is measured on held-out test data against external EEG/DRT labels, and the RF method's superiority is an empirical result, not a re-statement of its inputs.

full rationale

The paper's derivation chain is self-contained against the public MMDAP dataset with external ground-truth labels derived from EEG spectral features and DRT events. The proposed RF method is trained from vehicle-dynamics features and evaluated on a separate test split, so its reported accuracy (88%) and AUC (0.85) are measured outcomes rather than quantities defined by the model's own fit. The re-implemented prior methods (SvmA, SvmW, Lstm) are evaluated under the same common configuration C2 using the same labeling sources, so the observed discrepancies with reported values are empirical findings, not constructed equivalences. No fitted parameter is renamed as a prediction, and no load-bearing argument reduces to a self-citation. The paper's own threats-to-validity section discloses potential hyperparameter imbalance, feature bias toward RF (RF uses all feature sets), and individual differences; these are fairness and generalization risks, not circularity. The only notable protocol concern is that the 8:1:1 split (Sect. IV-C2) is described without explicit participant-level grouping, which could allow window-level leakage and inflate results; but even if true, that would be a data-protocol flaw affecting validity, not a circular derivation in which the output is equivalent to the input by construction. Therefore no circular step is identified.

Assumptions & free parameters 6 free parameters · 4 assumptions · 0 invented entities

The central comparison uses several externally adopted or hand-chosen conventions: EEG labeling thresholds from [4], C2 hyperparameters chosen by the authors, an 8:1:1 split with unstated grouping, and inferred re-implementation details from the original papers. None of these are derived from first principles; they are assumptions that the empirical ranking depends on.

free parameters (6)
  • EEG labeling quantile thresholds = 60% awake / 22.2% drowsy
    Adopted from [4]; chosen thresholds, not derived from the data.
  • C2 window size = 3 s
    Selected by the authors for the common configuration; no sensitivity analysis.
  • C2 overlap ratio = 50%
    Selected for the common configuration; affects sample count and independence.
  • C2 sampling rate = 60 Hz
    Common configuration choice; original methods used 10-60 Hz.
  • RF feature count (top-K) = not reported
    The number of features selected by ANOVA F-values is not stated; this parameter affects RF results and reproducibility.
  • Optuna hyperparameters for each model = not reported
    Search spaces and final hyperparameters are not disclosed, so tuning fairness cannot be assessed.
assumptions (4)
  • domain assumption EEG (theta+alpha)/beta power ratio is a valid objective drowsiness indicator, and quantile thresholds at 60% awake / 22.2% drowsy yield correct labels.
    Used in Sect. III-C to define ground truth for SvmA, SvmW, and RF. The indicator is based on [24] and thresholds on [4], but the paper does not validate them against behavioral outcomes in the MMDAP dataset.
  • domain assumption DRT event occurrence provides a valid binary drowsiness ground truth for the Lstm method.
    Used in Sect. III-D.3 and IV-C.1. The paper does not justify the mapping of event timing to drowsy/awake states or account for reaction-time variability.
  • domain assumption The 8:1:1 split creates independent train/validation/test sets without participant-level or temporal leakage.
    Foundation of the C2 evaluation in Sects. IV and V. The paper specifies only the ratio, not whether the split is by participant or by window, so independence is assumed.
  • domain assumption The re-implementations of [4], [5], and [6] faithfully capture the original methods despite missing details.
    The authors had to infer unspecified configurations (e.g., data split in [4], sampling rate in [5]). Errors in inference would change the reproducibility conclusions.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Comparison of Lightweight Methods for Vehicle Dynamics-Based Driver Drowsiness Detection." pith.science (2026). https://pith.science/paper/Q46F45TQ

@misc{pith2026250607014,
  author       = {Pith},
  title        = {Pith review of: Comparison of Lightweight Methods for Vehicle Dynamics-Based Driver Drowsiness Detection},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/Q46F45TQ}},
  note         = {Machine review of arXiv:2506.07014}
}
read the original abstract

Driver drowsiness detection (DDD) prevents road accidents caused by driver fatigue. Vehicle dynamics-based DDD has been proposed as a method that is both economical and high performance. However, there are concerns about the reliability of performance metrics and the reproducibility of many of the existing methods. For instance, some previous studies seem to have a data leakage issue among training and test datasets, and many do not openly provide the datasets they used. To this end, this paper aims to compare the performance of representative vehicle dynamics-based DDD methods under a transparent and fair framework that uses a public dataset. We first develop a framework for extracting features from an open dataset by Aygun et al. and performing DDD with lightweight ML models; the framework is carefully designed to support a variety of onfigurations. Second, we implement three existing representative methods and a concise random forest (RF)-based method in the framework. Finally, we report the results of experiments to verify the reproducibility and clarify the performance of DDD based on common metrics. Among the evaluated methods, the RF-based method achieved the highest accuracy of 88 %. Our findings imply the issues inherent in DDD methods developed in a non-standard manner, and demonstrate a high performance method implemented appropriately.

Figures

Figures reproduced from arXiv: 2506.07014 by the authors.

Figure 3
Figure 3. ML pipeline. wheel angle (SWA) θ (rad), steering wheel rate (SWR) ˙θ (rad/s), lateral velocity vx (m/s), longitudinal and lat￾eral accelerations ax, ay (m/s 2 ), and lane offset δ (m) [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗
Figure 1
Figure 1. The vehicle model [PITH_FULL_IMAGE:figures/full_fig_p004_1.png] view at source ↗
Figure 4
Figure 4. Example input signals. The sampling rate 10 Hz, window size 10 second without overlapping were set. 2) Discussion: The values we obtained were signifi￾cantly lower than RVs [PITH_FULL_IMAGE:figures/full_fig_p007_4.png] view at source ↗
Figures from the paper (1 more)
Figure 5
Figure 5. Figure 5: ROC curves. A. Results The results are shown in Tab. V and [PITH_FULL_IMAGE:figures/full_fig_p007_5.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

28 extracted references · 26 canonical work pages

  1. [1]

    Catastrophes, sleep, and public policy: consensus report,

    M.M.Mitler,M.A.Carskadon,C.A.Czeisler,W.C.Dement, D. F. Dinges, and R. C. Graeber, “Catastrophes, sleep, and public policy: consensus report,”Sleep, vol. 11, no. 1, pp. 100– 109, Feb. 1988, DOI: 10.1093/sleep/11.1.100

    work page doi:10.1093/sleep/11.1.100 1988

  2. [2]

    A Review of Recent Developments in Driver Drowsiness Detection Sys- tems,

    Y. Albadawi, M. Takruri, and M. Awad, “A Review of Recent Developments in Driver Drowsiness Detection Sys- tems,”Sensors, vol. 22, no. 5, article 2069, 2022, DOI: 10.3390/s22052069

    work page doi:10.3390/s22052069 2022

  3. [3]

    Challenges of Driver Drowsiness Prediction: The Remain- ing Steps to Implementation,

    E. Perkins, C. Sitaula, M. Burke, and F. Marzbanrad, “Challenges of Driver Drowsiness Prediction: The Remain- ing Steps to Implementation,”IEEE Transactions on Intel- ligent Vehicles, vol. 8, no. 2, pp. 1319–1338, Feb. 2023, DOI: 10.1109/TIV.2022.3224690

    work page arXiv 2023

  4. [4]

    Driver Drowsiness Detection Based on Steering Wheel Data Applying Adaptive Neuro-Fuzzy Feature Selection,

    S. Arefnezhad, S. Samiee, A. Eichberger, and A. Nahvi, “Driver Drowsiness Detection Based on Steering Wheel Data Applying Adaptive Neuro-Fuzzy Feature Selection,”Sensors, vol. 19, no. 4, article 943, 2019, DOI: 10.3390/s19040943

    work page doi:10.3390/s19040943 2019

  5. [5]

    Detecting Driver’s Drowsi- ness Using Multiwavelet Packet Energy Spectrum,

    S. F. Zhao, G. H. Xu, and T. Fei, “Detecting Driver’s Drowsi- ness Using Multiwavelet Packet Energy Spectrum,”Proceed- ings of the 2009 International Congress on Image and Signal Processing, pp. 1–5, 2009, DOI: 10.1109/CISP.2009.5301253

    work page arXiv 2009

  6. [6]

    Driver Distraction Detection Based on Vehicle Dynamics Using Nat- uralistic Driving Data,

    X. Wang, R. Xu, S. Zhang, Y. Zhuang, and Y. Wang, “Driver Distraction Detection Based on Vehicle Dynamics Using Nat- uralistic Driving Data,”Transportation Research Part C: Emerging Technologies, vol. 136, article 103561, 2022, DOI: 10.1016/j.trc.2022.103561

    work page arXiv 2022

  7. [7]

    Leakage and the Repro- ducibility Crisis in Machine-Learning-Based Science

    S. Kapoor, and A. Narayanan. “Leakage and the Repro- ducibility Crisis in Machine-Learning-Based Science.”Pat- terns, vol. 4, no. 9, article 100804, Sep. 2023, DOI: 10.1016/j.patter.2023.100804

    arXiv 2023

  8. [8]

    PPB-Emo: A Multimodal Psychological, Physiological and Behavioural Dataset for Human Emotions in Driving Tasks,

    W.Li,G.Guo,R.Tan,Y.Xing,G.Li,S.Li,etal.,“PPB-Emo: A Multimodal Psychological, Physiological and Behavioural Dataset for Human Emotions in Driving Tasks,”figshare, Collection, 2022, DOI: 10.6084/m9.figshare.c.5744171.v1

    work page doi:10.6084/m9.figshare.c.5744171.v1 2022

Show all 28 references
  1. [9]

    Dataset‘FULL’forDrowsi- ness Detection in Drivers,

    A.EichbergerandS.Arefnezhad,“Dataset‘FULL’forDrowsi- ness Detection in Drivers,”Graz University of Technology, 2022, DOI: 10.3217/8z09d-nrj27

    2022 doi

  2. [10]

    ApplyingDeepNeuralNetworksfor Multi-Level Classification of Driver Drowsiness Using Vehicle- Based Measures,

    S. Arefnezhad, S. Samiee, A. Eichberger, M. Frühwirth, C. Kaufmann,andE.Klotz,“ApplyingDeepNeuralNetworksfor Multi-Level Classification of Driver Drowsiness Using Vehicle- Based Measures,”Expert Systems with Applications, vol. 162, article 113778, 2020, DOI: 10.1016/j.eswa.2020.113778

    2020

  3. [11]

    Multi-modal Data Acquisition Platform for Behavioral Evaluation,

    A. Aygun, G. Blaney, Z. Haga, T. McWilliams, J. Mertens, J. P. de Ruiter, M. Scheutz, and N. Ward, “Multi-modal Data Acquisition Platform for Behavioral Evaluation,”Harvard Dataverse, version 2, 2024, DOI: 10.7910/DVN/HMZ5RG

    2024 doi

  4. [12]

    Young Novice Drivers’ Cognitive Distraction Detection: Comparing Support Vector Machines and Random Forest Model of Vehicle Control Be- havior,

    Q. Xue, X. Wang, Y. Li, and W. Guo, “Young Novice Drivers’ Cognitive Distraction Detection: Comparing Support Vector Machines and Random Forest Model of Vehicle Control Be- havior,”Sensors, vol. 23, no. 3, article 1345, Jan. 2023. DOI: 10.3390/s23031345

    2023 doi

  5. [13]

    Automatic Detection of Driver Fatigue Using Driving Operation Information for Transportation Safety,

    Z. Li, L. Chen, J. Peng, and Y. Wu, “Automatic Detection of Driver Fatigue Using Driving Operation Information for Transportation Safety,”Sensors, vol. 17, no. 6, article 1212, May 2017. DOI: 10.3390/s17061212

    2017 doi

  6. [14]

    SteeringWheelBehaviorBasedEstimationofFatigue,

    J. Krajewski, D. Sommer, U. Trutschel, D. Edwards, and M. Golz,“SteeringWheelBehaviorBasedEstimationofFatigue,” Proceedings of the Driving Assessment Conference, pp. 118– 124, 2009. DOI: 10.17077/drivingassessment.1311

    2009 doi

  7. [15]

    Evaluation of a Smart Algorithm for Commercial Vehicle Driver Drowsiness De- tection,

    A. Eskandarian and A. Mortazavi, “Evaluation of a Smart Algorithm for Commercial Vehicle Driver Drowsiness De- tection,”Proceedings of the 2007 IEEE Intelligent Vehi- cles Symposium, Istanbul, Turkey, pp. 553–559, 2007. DOI: 10.1109/IVS.2007.4290173

    2007

  8. [16]

    Steering in a Random Forest: Ensemble Learn- ing for Detecting Drowsiness-Related Lane Departures,

    A. D. McDonald, J. D. Lee, C. Schwarz, and T. L. Brown, “Steering in a Random Forest: Ensemble Learn- ing for Detecting Drowsiness-Related Lane Departures,”Hu- man Factors, vol. 56, no. 5, pp. 986–998, 2014. DOI: 10.1177/0018720813515272

    2014 doi

  9. [17]

    Real- Time Detection of Drowsiness Related Lane Departures Using Steering Wheel Angle,

    A. McDonald, C. Schwarz, J. Lee, and T. Brown, “Real- Time Detection of Drowsiness Related Lane Departures Using Steering Wheel Angle,”Proceedings of the Human Factors and Ergonomics Society Annual Meeting, vol. 56, pp. 210–215, 2012

    2012

  10. [18]

    OnlineDetectionof Driver Fatigue Using Steering Wheel Angles for Real Driving Conditions,

    Z.Li,S.E.Li,R.Li,B.Cheng,andJ.Shi,“OnlineDetectionof Driver Fatigue Using Steering Wheel Angles for Real Driving Conditions,”Sensors, vol. 17, no. 3, article 495, Mar. 2017. DOI: 10.3390/s17030495

    2017 doi

  11. [19]

    Particle Swarm Optimization of Feedforward Neural Networks for the Detection of Drowsy Driving,

    D. Sandberg and M. Wahde, “Particle Swarm Optimization of Feedforward Neural Networks for the Detection of Drowsy Driving,”Proceedings of the 2008 IEEE International Joint Conference on Neural Networks (IEEE World Congress on Computational Intelligence), Hong Kong, China, pp. 788–793,

    2008

  12. [20]

    Drowsiness Monitoring Based on Steering Wheel Status,

    M. Chai, S. Li, W. Sun, M. Guo, and M. Huang, “Drowsiness Monitoring Based on Steering Wheel Status,”Transportation Research Part D: Transport and Environment, vol. 66, pp. 95– 103, 2019. DOI: 10.1016/j.trd.2018.07.007

    2019 doi

  13. [21]

    Driver Cognitive Distraction Detection Using Driving Performance Measures,

    L. Jin, Q. Niu, H. Hou, H. Xian, Y. Wang, and D. Shi, “Driver Cognitive Distraction Detection Using Driving Performance Measures,”Discrete Dynamics in Nature and Society, vol. 2012, Art. no. 432634, Nov. 2012. DOI: 10.1155/2012/432634

    2012 doi

  14. [22]

    Estimating alertness from the EEG power spectrum,

    T. P. Jung, S. Makeig, M. Stensmo, and T. J. Sejnowski, “Estimating alertness from the EEG power spectrum,”IEEE Transactions on Biomedical Engineering, vol. 44, no. 1, pp. 60–69, Jan. 1997. DOI: 10.1109/10.553713

    1997 doi

  15. [23]

    Awareness of sleepiness when driving,

    J. A. Horne and S. D. Baulk, “Awareness of sleepiness when driving,”Psychophysiology, vol. 41, no. 1, pp. 161–165, Jan

  16. [24]

    Using EEG spectral components to assess algorithms for detecting fatigue [Part 1],

    B. Jap, S. Budi, S. Lal, P. Fischer, and E. Bekiaris, “Using EEG spectral components to assess algorithms for detecting fatigue [Part 1],”Expert Systems with Applications, vol. 36, pp. 2352-2359, 2009. DOI: 10.1016/j.eswa.2007.12.043

    2009 doi

  17. [25]

    Estimating Systemic Cognitive States from a Mixture of Physiological and Brain Signals,

    M. Scheutz, S. Aeron, A. Aygun, J. P. de Ruiter, S. Fantini, C. Fernandez, Z. Haga, T. Nguyen, and B. Lyu, “Estimating Systemic Cognitive States from a Mixture of Physiological and Brain Signals,”Topics in Cognitive Science, vol. 16, no. 3, pp. 485–526, Jul. 2024. DOI: 10.1111...

    2024 doi

  18. [26]

    Random Forests,

    L. Breiman, “Random Forests,”Machine Learning, vol. 45, no. 1, pp. 5–32, Oct. 2001. DOI: 10.1023/A:1010933404324

    2001 doi

  19. [2004]

    DOI: 10.1046/j.1469-8986.2003.00130.x

    2003

  20. [2008]

    DOI: 10.1109/IJCNN.2008.4633886

    2008

Pith tools

Reviewed August 7, 2026 · model on record in the stance chip above.