LLM Ensemble Fault Classification for Automotive HiL Validation

Andreas Rausch; Hamza Ouarrad; Mohammad Abboush

REVIEW 3 major objections 5 minor 38 references

LLM Ensemble Fault Classification for Automotive HiL Validation

T0 review · 3 major / 5 minor · reviewed 2026-08-12 · deepseek-v4-flash

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

Pith's one-line read A three-LLM ensemble with confidence-weighted voting beats every single model at classifying injected faults in automotive hardware-in-the-loop data.

desk verdict Plausible and useful in design, but the ensemble gain is not yet supported: the Top-3 set is selected on the same test scenarios whose improvement is reported. read the letter →

arxiv 2608.10710 v1 pith:D36AVB2L submitted 2026-08-11 cs.SE

Hamza Ouarrad , Mohammad Abboush , Andreas Rausch This is my paper · ORCID

classification cs.SE

keywords largelanguagemodelsensemblelearningfaultclassificationhardware-in-the-loopvalidationautomotivesoftwareexplainablediagnosisconfidence-weightedvotingtime-series

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 claims that a coordinated ensemble of three mid-sized open large language models can classify injected single faults in automotive hardware-in-the-loop (HiL) test recordings better than any individual model, reaching a scenario-averaged Top-1 accuracy of 0.917, macro F1 of 0.913, and Matthews correlation coefficient of 0.902, while also producing the most reliable confidence scores among tested fusion strategies. The authors argue this matters because current practice relies on manual review and rule-based anomaly detection that cannot name the fault type, while single-LLM diagnostics are prompt-sensitive, overconfident, and unstable. The proposed framework turns each faulty window into a compact statistical comparison against a time-aligned healthy reference, sends that evidence to three complementary LLMs, and fuses their structured outputs with confidence-weighted voting. A larger Top-5 ensemble does not improve results, which the paper reads as evidence that model complementarity matters more than ensemble size. If the claim holds, multi-LLM reasoning could give test engineers a reliable, calibrated, and explainable first pass over large volumes of HiL data.

What carries the argument

The load-bearing object is the compact evidence representation: instead of feeding raw time series, each 5-second faulty window is compared with the time-aligned healthy window and summarized per signal by statistical descriptors (mean difference, mean absolute deviation, maximum deviation, zero and saturation fractions, and a standardized deviation score). The prompt separates direct root-cause evidence from propagated support evidence, and the query is framed as a fault-only classification task with leakage-safe few-shot examples from a disjoint time region. The second mechanism is ensemble fusion: the same prompt is sent to three heterogeneous LLMs that each return a structured JSON with predicted class, confidence, ranked alternatives, and explanation, and the outputs are combined by confidence-weighted voting (with Borda count and reciprocal rank fusion as comparisons). This design is what lets the ensemble exploit model complementarity while keeping the input small enough for practical prompting and the output parseable for automated review.

What would settle it

Run the same pipeline on fault recordings that lack a paired healthy reference, or on a held-out HiL dataset with unseen driving profiles and fault classes; if Top-1 accuracy falls far below the reported 0.917 (or to chance level without the healthy baseline), the central empirical claim would be contradicted. A simpler check is to replace the healthy reference with a randomly chosen run of a different profile; the paper's premise predicts a large accuracy drop.

Watch

Extended reading notes

Core claim

The central discovery is that, on the evaluated gasoline-engine and electric-vehicle HiL fault-injection datasets across three driving settings and ten single-fault classes, the Top-3 ensemble of Mistral Small 24B, Qwen2.5 32B, and Phi-4 14B with confidence-weighted voting outperforms every single model on the main diagnostic metrics: 0.917 Top-1 accuracy, 0.913 macro F1, 0.902 MCC, and the lowest expected calibration error among fusion strategies. Borda rank aggregation and reciprocal rank fusion match or slightly beat the Top-2 accuracy but have much worse Brier scores and ECE, so the paper selects confidence-weighted voting as the final strategy. Adding Qwen2.5 14B and DeepSeek R1 Distill Qwen 32B to make a Top-5 ensemble slightly degrades accuracy and calibration, indicating diminishing returns from scale alone. The authors also show that the models' explanations refer to the expected root-cause signal before propagated evidence, which is the behavior a test engineer would want.

Load-bearing premise

Every faulty recording must have a time-aligned healthy recording of the same driving profile, because the framework only looks at deviations from that healthy run; without such a reference the pipeline cannot be applied at all.

Editorial extensions

If this is right

HiL validation teams could replace a large fraction of manual fault-label review with an LLM ensemble that names the fault, ranks alternatives, and points to the deciding signals, cutting validation turnaround time.
The 0.917 accuracy and 0.902 MCC are scenario-averaged across highway gasoline, urban gasoline, and highway EV settings, so the method is not tied to one propulsion architecture.
Because the ensemble returns ranked classes and explanations, it can function as a triage tool that flags the most confident diagnoses for automated handling and the less confident ones for engineer review, consistent with ISO 26262 traceability expectations.
Model size alone is not the driver: the quantized Llama 3.1 70B underperforms smaller models, and the Top-5 ensemble underperforms the Top-3, so the practical guidance is to select a small complementary set of models rather than the largest available.
The confidence scores of the Top-3 ensemble are calibrated well enough that downstream systems could threshold on them: the Brier score of 0.153 and ECE of 0.058 are the best among fusion strategies.

Reading between the lines

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

A direct testable extension is to ablate the healthy reference: feed the same prompts without the time-aligned baseline and measure the accuracy drop; the paper's premise predicts a large collapse, which would confirm how dependent the result is on that design choice.
The compact-evidence prompting style could transfer to other XiL domains (e.g., brake-by-wire, ADAS sensor validation) where a healthy reference run is available, though the 0.917 figure is specific to these two platforms and ten fault classes and should not be extrapolated.
The complementarity-over-size result suggests ensemble selection could be done by measuring pairwise disagreement among candidate LLMs on a small validation set, rather than by adding models blindly; the paper does not propose such a selection criterion.
The explanation outputs might be used to build a feedback loop that catches misclassifications: if a top-ranked class's explanation cites only support signals and never the expected root signal, an engineer could flag that prompt for inspection; this is a use the paper gestures at but does not formalize.

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, and a circularity audit.

Desk Editor's Note

read the letter

Quick take: the claimed Top-3 ensemble advantage over the best single model is real in their numbers, but the evaluation as reported cannot support it, because the three ensemble members were selected from the same scenarios used to report the final gain. The +0.014 Top-1 and +0.015 MCC improvement over Mistral Small is exactly the kind of gap that post hoc model selection on the test set can manufacture.

What is actually new: as far as the cited literature goes, this is the first coordinated multi-LLM ensemble for sensor-level fault classification in HiL validation. The design has several good properties: instead of prompting raw time-series, they compute compact deviation statistics against a time-aligned healthy reference; the few-shot examples come from a separate time interval with an embargo region; and the structured JSON output with ranked classes and explanations is a sensible way to get not just labels but a justification an engineer can check. The explanation examples in Table 4 look like the model really is keying on the root-cause signal first, which is the right behavior for this use case.

The soft spots are real. The main one is the load-bearing evaluation flaw: Section 5.2 says the Top-3 set was chosen because those models had the strongest single-model results on the same scenario-averaged windows that produce Table 3. There is no held-out model selection or pre-registered model set. Second, the ensemble gain is small and unevenly distributed: in Table 2, two of the three scenarios are already perfect for every fusion method, so the entire ensemble value rests on the Gas-Hwy scenario with Top-1=0.750. Third, all numbers are point estimates—no confidence intervals, no per-window counts, no repeated runs, no non-LLM baseline. Fourth, no data, prompts, or code are released. The healthy-reference requirement is a genuine scope condition, but it is explicitly assumed from Section 3.1 and does not threaten the internal validity of the comparison the paper actually runs.

The citation pattern looks fine; the related work is relevant and recent, and the authors do not oversell prior work. There is a minor text corruption in Section 5.2 ("reaches a highered voting achieves...") that should be cleaned up.

Who this is for: researchers working on LLM-based diagnostics and anyone building evaluation pipelines for time-series classification. It deserves peer review, but the revision needs to fix the selection problem, add uncertainty quantification, include a classical baseline, and release artifacts. I would send it to referees rather than desk-reject it.

Referee Report

3 major / 5 minor

Summary. The paper proposes an explainable multi-LLM ensemble framework for sensor-level fault classification in automotive hardware-in-the-loop (HiL) validation. The pipeline converts windowed multivariate signals into compact statistical evidence by comparing faulty recordings with time-aligned healthy references, builds structured few-shot prompts, and queries several open-weight LLMs. The individual predictions and confidences are fused with three strategies (confidence-weighted voting, Borda count, reciprocal rank fusion). The authors evaluate seven LLMs on three driving scenarios from two HiL systems (gasoline-engine highway, gasoline-engine urban, electric-vehicle highway) with ten single-fault classes. They report that a Top-3 ensemble of Mistral Small 24B, Qwen2.5 32B, and Phi-4 14B with confidence-weighted voting achieves scenario-averaged Top-1 accuracy 0.917, macro F1 0.913, and MCC 0.902, outperforming every single model, with the best calibration among the tested ensemble strategies. A Top-5 ensemble does not improve over the Top-3 configuration.

Significance. If the reported results are reliable, the paper would provide a useful demonstration that coordinated multi-LLM reasoning with compact, reference-based evidence can outperform single LLMs on structured diagnostic classification while generating engineer-readable explanations. The study is one of the first to apply LLM ensembles to sensor-level fault diagnosis in automotive HiL data, and it contributes a concrete prompt-engineering recipe and a leakage-aware train/embargo/test split. The runtime table and the explicit comparison of fusion strategies are also valuable. However, the significance is tempered by the evaluation design: the ensemble members and the final fusion strategy are selected using the same evaluation scenarios on which the improvement is reported, and the statistical evidence is thin (three scenarios, no confidence intervals, no repeated runs). The framework's dependence on a healthy reference run is a real scope limitation, though the paper states this assumption explicitly.

major comments (3)

[Section 5.2, Table 1 vs. Table 3] The Top-3 model selection and the choice of confidence-weighted voting are made after inspecting the same scenario-averaged results that are then reported as the final ensemble gain. Section 5.2 states that the three models were 'selected' because they achieved the strongest single-model results, and the fusion strategy is chosen using the same Table 2/3 numbers. No held-out model-selection set, embargo-based selection, or pre-registered model set is described. With seven candidate models and a reported improvement of only +0.014 Top-1 accuracy and +0.015 MCC over the best single model, post hoc selection on the test set can plausibly manufacture the observed difference. The authors should either perform model selection on a separate validation split (e.g., the 25–30 s embargo region) or re-frame the single-model comparison and ensemble gain as exploratory and report selection-corrected estimates.
[Section 5.2, Tables 2 and 3] The scenario-averaged ensemble gain rests almost entirely on the Gas-Hwy scenario, because Gas-Urb and EV-Hwy achieve perfect 1.000 scores for every fusion method. Yet the paper reports no window counts, no confidence intervals, and no repeated runs to bound the variance of the 0.917 Top-1 estimate. LLM outputs are stochastic, and without temperature/seed details or multiple inference runs, the +0.014 improvement over Mistral 24B may be within run-to-run noise. The authors should report per-window sample sizes, bootstrap confidence intervals, and per-scenario error bars, and should explicitly acknowledge that the effective evidence for the ensemble advantage comes from one difficult scenario.
[Section 3.1 and Section 4.3] The entire framework depends on the availability of a time-aligned healthy reference run for each faulty recording under the same driving profile. The paper states this as an assumption and uses it to compute deviation evidence, but it provides no evidence about how often such references exist in production HiL campaigns or how the method behaves when a reference is missing or misaligned. Since this is a load-bearing scope condition, the authors should state it as an explicit limitation and, ideally, provide a sensitivity analysis with perturbed or missing references.

minor comments (5)

[Section 5.2, paragraph after Table 3] The sentence 'Although Borda rank aggregation and reciprocal rank fusion reach a highered voting achieves 0.917 Top-1 accuracy, 0.913 macro F1, and Top-2 accuracy of 0.972' contains a text corruption and is not readable; it should be rewritten.
[Table 3, footnote] The table footnote says 'The dagger marks the final selected ensemble strategy', but no dagger appears in the table; either add the dagger to the CW vote row or change the footnote.
[Table 2, formatting] Several rows lack spaces between method names and values (e.g., 'Borda1.000', 'RRF1.000'), making the table harder to read.
[Table 4, formatting] In the 'True / predicted' row, the format 'predictedBAT_HV_Gain_3' and 'predictedAccPedal_Gain_5' is missing a separating space; this should be fixed.
[Section 5.2, calibration comparison] The paper emphasizes that confidence-weighted voting provides the best calibration among ensemble strategies, but the CW Brier score of 0.153 is notably worse than the single-model Mistral 24B Brier of 0.102 (Table 1). Since the introduction motivates ensembles as improving 'confidence reliability', the authors should explicitly compare ensemble calibration with the single-model baseline and temper any implication that the ensemble improves calibration overall.

Circularity Check

1 steps flagged · score 3.0 of 10

The Top-3 model set and the confidence-weighted fusion rule are selected on the same scenario-averaged test windows whose improvement is then reported, so the headline gain is partly an in-sample selection artifact rather than an independent prediction.

fitted input called prediction [Section 5.2 (Tables 1-3); abstract]
"After evaluating the individual LLMs, the three strongest and most complementary models were selected for ensemble fusion: Mistral Small 24B, Qwen2.5 32B, and Phi-4 14B. These models were chosen because they achieved the strongest single-model results in terms of Top-1 accuracy, macro F1, MCC, and Top-2 ranking capability. ... The scenario-averaged results in Table 3 confirm that the Top-3 ensemble improves the overall diagnostic performance."

All performance numbers are reported as scenario-averaged values across the same three settings. Table 1 is the basis for choosing the Top-3 members; Table 3 is the basis for the reported ensemble gain. Both tables are computed from the same 30-45 s evaluation windows. The ensemble configuration is therefore fitted to the exact data that are then cited as evidence that the ensemble 'improves' diagnostic performance. Confidence-weighted voting is also chosen because it 'provides the best balance' in the same Table 3. The abstract's 'improving the scenario-averaged results to 0.917' is thus an in-sample report, not an out-of-sample prediction. The gain is not forced by an equation, so this is partial selection-induced circularity rather than a definitional reduction.

full rationale

The paper makes no mathematical derivation, so there is no equation-level self-definition and no imported uniqueness theorem. The time-aligned healthy-reference requirement is an explicit scope assumption in Section 3.1, not a circular step, and the related-work citations, including two co-authored papers, are not used to justify the central performance claim. The only substantive circularity concern is experimental: Section 5.2 selects both the Top-3 model set and the final fusion strategy using the same scenario-averaged evaluation windows that later appear as the reported improvement. This is a test-set selection problem rather than a proof-by-definition: the ensemble score of 0.917 is not mathematically forced to exceed the best single model, and a held-out validation split would have made the comparison clean. I therefore assign a moderate-low score: the headline advantage is partly an in-sample artifact, but the framework itself is a genuine empirical proposal whose claims could be falsified on new data.

Assumptions & free parameters 3 free parameters · 5 assumptions · 0 invented entities

The framework depends on the availability of time-aligned healthy references, a fixed ten-class single-fault taxonomy, and the sufficiency of condensed statistical descriptors. The paper also selects its three ensemble models using the same scenarios on which final results are reported. No new physical entities or conserved quantities are introduced.

free parameters (3)

Window length and step = 5 s window, 1 s step
Manual segmentation choice with no ablation study. The central results depend on this unexamined setting.
Active fault interval and split boundaries = 15-45 s; few-shot 15-25 s, embargo 25-30 s, test 30-45 s
Transient phases are excluded by hand. The test span is only 15 s per recording, which may explain perfect scores in two of three scenarios.
Top-3 model selection = Mistral Small 24B, Qwen2.5 32B, Phi-4 14B
Models were chosen by their best performance on the same evaluation scenarios used for the final ensemble results, a post-hoc selection on the test set.

assumptions (5)

domain assumption For each faulty recording, a time-aligned healthy reference run under the same driving profile exists and is used as the baseline.
Section 3.1 makes deviations from the healthy run the only input. Without such a reference, the framework is undefined.
domain assumption Each test window contains exactly one of the ten known single-fault classes; concurrent faults and unknown faults are out of scope.
Section 3.3 frames the task as fault-only single-fault classification, and Section 6 lists concurrent faults as future work.
domain assumption Compact statistical descriptors preserve enough information for fault classification.
Section 3.2 replaces raw signals with descriptors such as mean difference, MAD, max deviation, and standardized deviation. No analysis shows that this representation is sufficient for the fault taxonomy.
domain assumption LLM outputs are valid structured JSON and reported confidences are meaningful probability-like scores.
Section 3.3 requires structured JSON output. Parse success rates are not reported, and calibration is only checked via aggregate Brier and ECE.
domain assumption Scenario averages across the three settings are equally weighted and representative of HiL validation practice.
Tables 2 and 3 treat gasoline highway, gasoline urban, and electric highway as one average each, with no variance or population justification.

how reviews work

0 comments

Cite this review

Pith. "Pith review of LLM Ensemble Fault Classification for Automotive HiL Validation." pith.science (2026). https://pith.science/paper/D36AVB2L

@misc{pith2026260810710,
  author       = {Pith},
  title        = {Pith review of: LLM Ensemble Fault Classification for Automotive HiL Validation},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/D36AVB2L}},
  note         = {Machine review of arXiv:2608.10710}
}

read the original abstract

Automotive HiL validation generates large multivariate test recordings whose analysis remains challenging due to manual review effort, rule-based limitations, and the need for explainable diagnostic decisions. Recent machine-learning and deep-learning approaches have improved fault diagnosis, but they often require large labelled datasets, generalise poorly across operating conditions, and provide limited insight into their predictions. This paper proposes an explainable multi-LLM ensemble framework for sensor-level fault classification in automotive validation. The framework uses compact evidence representations of fault-injection recordings and combines the outputs of heterogeneous large language models to improve diagnostic robustness, ranking quality, confidence reliability, and interpretability. The approach is evaluated on gasoline-engine and electric-vehicle HiL systems across three driving settings and ten single-fault classes. Among the individual models, Mistral Small~24B provides the strongest overall single-model trade-off, achieving 0.903 Top-1 accuracy, 0.887 MCC, and the lowest Brier score of 0.102. The final Top-3 ensemble combines Mistral Small~24B, Qwen2.5~32B, and Phi-4~14B using confidence-weighted voting, improving the scenario-averaged results to 0.917 Top-1 accuracy, 0.913 macro F1, and 0.902 MCC, while also providing the best calibration among the tested ensemble strategies. A Top-5 ensemble does not improve over the Top-3 configuration, indicating that model complementarity is more important than ensemble size. The results show that coordinated multi-LLM reasoning can support robust, calibrated, and engineer-interpretable fault classification for automotive HiL validation.

Figures

Figures reproduced from arXiv: 2608.10710 by the authors.

**Figure 3.** Comparison of Top-3 and Top-5 confidence [PITH_FULL_IMAGE:figures/full_fig_p006_3.png] view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

38 extracted references · 35 canonical work pages

[1]

Mohammad Abboush, Christoph Knieke, and Andreas Rausch. 2023. Intelligent identification of simultaneous faults of automotive software systems under noisy and imbalanced data using ensemble LSTM and random forest.IEEE Access11 (2023), 140022–140040

work page 2023
[2]

Mohammad Abboush, Christoph Knieke, and Andreas Rausch. 2024. A virtual testing framework for real-time validation of automotive software systems based on hardware in the loop and fault injection.Sensors24, 12 (2024), 3733

work page 2024
[3]

Mohammed Nasser Al-Andoli, Shing Chiang Tan, Kok Swee Sim, Manjeevan Seera, and Chee Peng Lim. 2023. A parallel ensemble learning model for fault detection and diagnosis of industrial machinery.IEEE Access11 (2023), 39866– 39878

work page 2023
[4]

Usman Ali, Umer Ramzan, Waqas Ali, and Khaled Ali Al-Jaafari. 2025. An improved fault diagnosis strategy for induction motors using weighted probability ensemble deep learning.IEEE Access(2025)

work page 2025
[5]

Tahir Anwar, Chaoxu Mu, Muhammad Zain Yousaf, Wajid Khan, Saqib Khalid, Ahmad O Hourani, and Ievgen Zaitsev. 2025. Robust fault detection and clas- sification in power transmission lines via ensemble machine learning models. Scientific Reports15, 1 (2025), 2549

work page 2025
[6]

Sihem Baccari, Mohamed Hadded, Hakim Ghazzai, Haifa Touati, and Mourad Elhadef. 2024. Anomaly detection in connected and autonomous vehicles: A survey, analysis, and research challenges.IEEE access12 (2024), 19250–19276

2024
[7]

Yushen Chen, Guan Qiang, Ximing Xie, Boyu Wang, Donghong Cai, Muham- mad Diyan, Tianlong Liu, and Fang Fang. 2026. Personalized Vehicular Health Diagnosis via Large Language Model.IEEE Communications Standards Magazine (2026)

work page 2026
[8]

Xinyu Dai and Zhiwei Gao. 2013. From Model, Signal to Knowledge: A Data- Driven Perspective of Fault Detection and Diagnosis.IEEE Transactions on Industrial Informatics9, 4 (2013), 2226–2238

work page 2013

Show all 38 references

[9]

Rakesh Das, Henry Griffith, and Heena Rathore. 2025. Llm-based fault detection in connected vehicle time-series data. InProceedings of the 2025 ACM Workshop on Secure and Trustworthy Cyber-physical Systems. 13–22

2025
[10]

Simon Diemert and Jens H Weber. 2023. Can large language models assist in hazard analysis?. InInternational Conference on Computer Safety, Reliability, and Security. Springer, 410–422

2023
[11]

Chao Feng, Zihan Liu, Siddhant Gupta, Gongpei Cui, Jan von der Assen, and Burkhard Stiller. 2025. Smarter, not Bigger: Fine-Tuned RAG-Enhanced LLMs for Automotive HIL Testing.arXiv preprint arXiv:2511.22584(2025)

2025
[12]

Dacian Goina, Eduard Hogea, and George Maties. 2024. Enhanced Anomaly Detection in Automotive Systems Using SAAD: Statistical Aggregated Anom- aly Detection. In2024 26th International Symposium on Symbolic and Numeric Algorithms for Scientific Computing (SYNASC). IEEE, 233–241

2024
[13]

Qingnan Huang, Benhao Liang, Xisheng Dai, Shan Su, and Enze Zhang. 2024. Unmanned aerial vehicle fault diagnosis based on ensemble deep learning model. Measurement Science and Technology35, 4 (2024), 046205

2024
[14]

International Organization for Standardization. 2018. ISO 26262-10:2018-Road vehicles—Functional safety—Part 10: Guideline on ISO 26262. https://www.iso. org/standard/68392.html. Accessed on: Feb. 14, 2026, [Online]

2018
[15]

Udeme Ibanga Inyang, Ivan Petrunin, and Ian Jennions. 2023. Diagnosis of multiple faults in rotating machinery using ensemble learning.Sensors23, 2 (2023), 1005

2023
[16]

Halid Kaplan, Kambiz Tehrani, and Mo Jamshidi. 2021. A fault diagnosis design based on deep learning approach for electric vehicle applications.Energies14, 20 (2021), 6599

2021
[17]

Lars Klitzke, Carsten Koch, Andreas Haja, and Frank Köster. 2019. Real-world Test Drive Vehicle Data Management System for Validation of Automated Driving Systems.. InVEHITS. 171–180

2019
[18]

Jeong-Woo Lee, Ki-Yong Choi, and Jung-Won Lee. 2017. Collecting big data from automotive ECUs beyond the CAN bandwidth for fault visualization.Mobile Information Systems2017, 1 (2017), 4395070

2017
[19]

Liang Lei, Weibin Li, Shiwei Zhang, Changyuan Wu, and Hongxiang Yu. 2025. Research progress on data-driven industrial fault diagnosis methods.Sensors25, 9 (2025), 2952

2025
[20]

Lin Lin, Sihao Zhang, Song Fu, and Yikun Liu. 2025. FD-LLM: Large language model for fault diagnosis of complex equipment.Advanced Engineering Infor- matics65 (2025), 103208

2025
[21]

Ehtisham Lodhi, Fei-Yue Wang, Gang Xiong, Lingjian Zhu, Tariku Sinshaw Tamir, Waheed Ur Rehman, and M Adil Khan. 2023. A novel deep stack-based ensemble learning approach for fault detection and classification in photovoltaic arrays. Remote Sensing15, 5 (2023), 1277

2023
[22]

You Lu, Yifan Tian, Yuyang Bi, Bihuan Chen, and Xin Peng. 2024. Diavio: Llm- empowered diagnosis of safety violations in ads simulation testing. InProceedings of the 33rd ACM SIGSOFT International Symposium on Software Testing and Anal- ysis. 376–388

2024
[23]

Yashashree Mahale, Shrikrishna Kolhar, and Anjali S More. 2025. Automated vehicle fault diagnosis and report generation using hybrid machine learning with multi-step RAG approach.Discover Computing28, 1 (2025), 283. MoDeVVa ’26, October 04–06, 2026, Málaga, Spain Ouarrad et al

2025
[24]

Rafia Nishat Toma and Jong-Myon Kim. 2020. Bearing fault classification of in- duction motors using discrete wavelet transform and ensemble machine learning algorithms.Applied Sciences10, 15 (2020), 5251

2020
[25]

Fannia Pacheco, Alin Drimus, Lars Duggen, Mariela Cerrada, Diego Cabrera, and René-Vinicio Sanchez. 2022. Deep ensemble-based classifier for transfer learning in rotating machinery fault diagnosis.IEEE Access10 (2022), 29778–29787

2022
[26]

John Pavlopoulos, Alv Romell, Jacob Curman, Olof Steinert, Tony Lindgren, Markus Borg, and Korbinian Randl. 2024. Automotive fault nowcasting with machine learning and natural language processing.Machine learning113, 2 (2024), 843–861

2024
[27]

Wojciech Pietrowski, Magdalena Puskarczyk, and Jan Szymenderski. 2024. Fault Detection Methods for Electric Power Steering System Using Hardware in the Loop Simulation.Energies17, 14 (2024), 3486

2024
[28]

Radhika Raveendran, KB Devika, and Shankar C Subramanian. 2020. Brake fault identification and fault-tolerant directional stability control of heavy road vehicles.IEEE Access8 (2020), 169229–169246

2020
[29]

Saeid Safavi, Mohammad Amin Safavi, Hossein Hamid, and Saber Fallah. 2021. Multi-sensor fault detection, identification, isolation and health forecasting for autonomous vehicles.Sensors21, 7 (2021), 2547

2021
[30]

2013.Detecting anomalies in multivariate time series from automotive systems

Andreas Theissler. 2013.Detecting anomalies in multivariate time series from automotive systems. Ph. D. Dissertation. Brunel University School of Engineering and Design PhD Theses

2013
[31]

Andreas Theissler. 2017. Detecting known and unknown faults in automotive systems using ensemble-based anomaly detection.Knowledge-Based Systems123 (2017), 163–173

2017
[32]

Andreas Theissler and Ian Dear. 2013. An anomaly detection approach to detect unexpected faults in recordings from test drives. InProceedings of the W ASET International Conference on Vehicular Electronics and Safety, Vol. 7. 195–198

2013
[33]

Ghizlane Tibba, Christoph Malz, Christoph Stoermer, Natarajan Nagarajan, Li- cong Zhang, and Samarjit Chakraborty. 2016. Testing automotive embedded systems under X-in-the-loop setups. In2016 IEEE/ACM International Conference on Computer-Aided Design (ICCAD). IEEE, 1–8

2016
[34]

Shuai Wang, Yinan Yu, Robert Feldt, and Dhasarathy Parthasarathy. 2025. Au- tomating a complete software test process using llms: An automotive case study. arXiv preprint arXiv:2502.04008(2025)

2025 arXiv
[35]

Haiyu Zhang, Yinghui Zhao, Boyu Sun, Yaqi Wu, Zetian Fu, and Xinqing Xiao
[36]

Yuzhong Zhang, Songyang Zhang, and Venkata Dinavahi. 2025. A survey of machine learning applications in advanced transportation systems: Trends, tech- niques, and future directions.ETransportation24 (2025), 100417

2025
[37]

Shuwen Zheng, Kai Pan, Jie Liu, and Yunxia Chen. 2024. Empirical study on fine-tuning pre-trained large language models for fault diagnosis of complex systems.Reliability Engineering & System Safety252 (2024), 110382

2024
[2025]

Large language model based intelligent fault information retrieval system for new energy vehicles.Applied Sciences15, 7 (2025), 4034

2025

Pith tools

Reviewed August 12, 2026 · model on record in the stance chip above.

[1] [1]

Mohammad Abboush, Christoph Knieke, and Andreas Rausch. 2023. Intelligent identification of simultaneous faults of automotive software systems under noisy and imbalanced data using ensemble LSTM and random forest.IEEE Access11 (2023), 140022–140040

work page 2023

[2] [2]

Mohammad Abboush, Christoph Knieke, and Andreas Rausch. 2024. A virtual testing framework for real-time validation of automotive software systems based on hardware in the loop and fault injection.Sensors24, 12 (2024), 3733

work page 2024

[3] [3]

Mohammed Nasser Al-Andoli, Shing Chiang Tan, Kok Swee Sim, Manjeevan Seera, and Chee Peng Lim. 2023. A parallel ensemble learning model for fault detection and diagnosis of industrial machinery.IEEE Access11 (2023), 39866– 39878

work page 2023

[4] [4]

Usman Ali, Umer Ramzan, Waqas Ali, and Khaled Ali Al-Jaafari. 2025. An improved fault diagnosis strategy for induction motors using weighted probability ensemble deep learning.IEEE Access(2025)

work page 2025

[5] [5]

Tahir Anwar, Chaoxu Mu, Muhammad Zain Yousaf, Wajid Khan, Saqib Khalid, Ahmad O Hourani, and Ievgen Zaitsev. 2025. Robust fault detection and clas- sification in power transmission lines via ensemble machine learning models. Scientific Reports15, 1 (2025), 2549

work page 2025

[6] [6]

Sihem Baccari, Mohamed Hadded, Hakim Ghazzai, Haifa Touati, and Mourad Elhadef. 2024. Anomaly detection in connected and autonomous vehicles: A survey, analysis, and research challenges.IEEE access12 (2024), 19250–19276

2024

[7] [7]

Yushen Chen, Guan Qiang, Ximing Xie, Boyu Wang, Donghong Cai, Muham- mad Diyan, Tianlong Liu, and Fang Fang. 2026. Personalized Vehicular Health Diagnosis via Large Language Model.IEEE Communications Standards Magazine (2026)

work page 2026

[8] [8]

Xinyu Dai and Zhiwei Gao. 2013. From Model, Signal to Knowledge: A Data- Driven Perspective of Fault Detection and Diagnosis.IEEE Transactions on Industrial Informatics9, 4 (2013), 2226–2238

work page 2013