Pith. sign in

REVIEW 3 major objections 2 minor 48 references

Natural-language driving preferences can be translated into distinguishable lane-change behaviors in the Apollo automated driving stack using an LLM framework.

Reviewed by Pith at T0; open to challenge. T0 means a machine referee read the full paper against a public rubric. the ladder, T0–T4 →

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

T0 review · grok-4.3

2026-07-02 19:10 UTC pith:V7MJJFE6

load-bearing objection The paper shows a practical LLM+RAG setup that maps implicit commands to clustered Apollo lane-change parameters with released code, but the clustering details and quantitative validation are thin. the 3 major comments →

arxiv 2606.31483 v2 pith:V7MJJFE6 submitted 2026-06-30 cs.RO

A Large-Language-Model Supported Personalized Driving Framework for Lane Change in Highway Scenarios

Dong Bi , Yongqi Zhao , Paul Kovacevic , Tomislav Mihalj , Ji Zhou , Jiayuan Gong , Arno Eichberger This is my paper
classification cs.RO
keywords personalized drivinglane changelarge language modelApollo automated drivingretrieval-augmented generationhighway scenariosdriving stylesnatural language commands
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved

The pith

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

The paper proposes a framework that uses large language models to map user commands to planning parameters for lane changes on highways, categorized into aggressive, normal, and conservative styles. It evaluates candidate parameters by simulating behaviors, then clusters and ranks them by style intensity to create executable sets. For interpreting commands, especially implicit ones, it employs retrieval-augmented generation to improve accuracy. A sympathetic reader would care because this could make automated driving systems more acceptable by allowing natural interaction rather than technical inputs.

Core claim

The framework establishes a mapping from natural-language driving commands to executable planning parameters in the Apollo stack by constructing style-specific parameter sets through evaluation, clustering, and ranking of lane-change behaviors. Retrieval-augmented generation with the LLM enhances the interpretation of both explicit and implicit user preferences, leading to personalized and distinguishable driving behaviors.

What carries the argument

The style-specific parameter sets derived from clustering and style-intensity ranking of candidate planning parameters, combined with RAG for command interpretation.

Load-bearing premise

That candidate planning parameters can be clustered and ranked by style intensity to produce truly distinguishable and style-specific executable sets in the Apollo stack.

What would settle it

If experiments show that the lane-change behaviors from the different style parameter sets are not distinguishable by observers or metrics, the claim would be falsified.

Watch this falsifier. Get emailed when new claim-graph text bears on it.

Share X Bluesky LinkedIn Reddit HN

If this is right

  • The derived parameter sets generate distinguishable personalized lane-change behaviors for aggressive, normal, and conservative styles.
  • RAG consistently improves preference interpretation, particularly for implicit commands.
  • Integration of LLM-based natural-language interaction with Apollo supports personalized lane-change behavior generation.

Where Pith is reading between the lines

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

  • If the framework generalizes, it could extend to other driving maneuvers beyond lane changes.
  • User acceptance of automated vehicles might increase with natural language interfaces.
  • Testing in real-world conditions could validate the simulation-based clustering.

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 proposes an LLM-supported framework for personalized highway lane-change driving in the Apollo stack. Natural-language commands (explicit and implicit) are mapped to executable planning parameters grouped into aggressive, normal, and conservative styles. The mapping is built by evaluating candidate parameter sets on resulting lane-change behaviors, then constructing style-specific sets via clustering and style-intensity ranking; a separate retrieval dataset enables RAG to improve LLM interpretation of preferences. Experiments are reported to show that the derived parameter sets produce distinguishable behaviors and that RAG improves interpretation, especially for implicit commands. Code and datasets are released.

Significance. If the distinguishability and RAG results hold under rigorous validation, the work would offer a practical route to natural-language personalization of production-grade AV stacks, addressing a gap between high-level user preferences and low-level planning parameters. The open release of code and datasets is a clear strength that supports reproducibility and extension.

major comments (3)
  1. [Framework description / parameter-set construction] The construction of style-specific parameter sets (described after the framework overview) does not specify the behavior feature vector (e.g., completion time, peak lateral acceleration, lane-change duration), distance metric, clustering algorithm, or any statistical test used to confirm separation among aggressive/normal/conservative clusters. Without these, the central claim that the sets generate “distinguishable personalized lane-change behaviors” cannot be evaluated.
  2. [Experimental results] The experimental results paragraph (and abstract) states positive outcomes on distinguishability and RAG improvement but reports no quantitative metrics, baselines, statistical tests, sample sizes, or exclusion criteria. This absence directly undermines assessment of the “distinguishable” and “consistently improves” claims.
  3. [RAG-supported command interpretation] The RAG evaluation relies on an independently constructed retrieval dataset, yet no details are given on how preference-interpretation accuracy is measured (e.g., human ratings, execution success rate in Apollo, inter-rater agreement) or on the comparison condition without RAG.
minor comments (2)
  1. Notation for the three driving styles is introduced inconsistently (sometimes “aggressive/normal/conservative,” sometimes abbreviated); a single table or definition section would improve clarity.
  2. The abstract claims “experimental results show…” but the full results section should include at least one table or figure with the actual performance numbers to support the textual claims.

Simulated Author's Rebuttal

3 responses · 0 unresolved

We thank the referee for the constructive and detailed feedback. We address each major comment below and indicate the revisions that will be incorporated into the manuscript.

read point-by-point responses

  1. Referee: [Framework description / parameter-set construction] The construction of style-specific parameter sets (described after the framework overview) does not specify the behavior feature vector (e.g., completion time, peak lateral acceleration, lane-change duration), distance metric, clustering algorithm, or any statistical test used to confirm separation among aggressive/normal/conservative clusters. Without these, the central claim that the sets generate “distinguishable personalized lane-change behaviors” cannot be evaluated.

    Authors: We agree that these methodological details are necessary to substantiate the distinguishability claim. The revised manuscript will explicitly describe the behavior feature vector (including completion time, peak lateral acceleration, and lane-change duration), the distance metric, the clustering algorithm, and the statistical tests used to confirm separation among the clusters. revision: yes

  2. Referee: [Experimental results] The experimental results paragraph (and abstract) states positive outcomes on distinguishability and RAG improvement but reports no quantitative metrics, baselines, statistical tests, sample sizes, or exclusion criteria. This absence directly undermines assessment of the “distinguishable” and “consistently improves” claims.

    Authors: We acknowledge that the current experimental results section lacks the requested quantitative details. The revision will add specific quantitative metrics, any baselines employed, statistical test outcomes, sample sizes, and exclusion criteria to support the reported claims. revision: yes

  3. Referee: [RAG-supported command interpretation] The RAG evaluation relies on an independently constructed retrieval dataset, yet no details are given on how preference-interpretation accuracy is measured (e.g., human ratings, execution success rate in Apollo, inter-rater agreement) or on the comparison condition without RAG.

    Authors: We will revise the RAG evaluation section to specify how accuracy is measured (including human ratings, Apollo execution success rates, and inter-rater agreement) and to present the direct comparison against the non-RAG baseline. revision: yes

Circularity Check

0 steps flagged

No significant circularity; parameter sets derived from external Apollo evaluations and independent RAG dataset

full rationale

The framework evaluates candidate planning parameters on lane-change behaviors executed in Apollo, clusters them by style intensity, and builds a separate retrieval dataset for RAG-based command interpretation. These steps rely on external simulation outputs and an independently constructed dataset rather than reducing results to quantities defined by the paper's own fitted parameters or self-citations. The distinguishability claim is supported by post-derivation execution tests, not tautological by construction. Minor self-citation risk is possible but not load-bearing based on provided text.

Axiom & Free-Parameter Ledger

1 free parameters · 1 axioms · 0 invented entities

The central claim rests on the Apollo stack correctly executing mapped parameters and on the constructed retrieval dataset being representative for RAG; no new entities postulated.

free parameters (1)
  • style-specific planning parameter sets
    Derived through evaluation of candidates, clustering, and style-intensity ranking to match aggressive, normal, and conservative behaviors.
axioms (1)
  • domain assumption The Apollo automated driving stack can translate the selected planning parameters into executable lane-change trajectories.
    Framework maps to Apollo parameters and relies on its execution for behavior generation.

pith-pipeline@v0.9.1-grok · 5761 in / 1193 out tokens · 30254 ms · 2026-07-02T19:10:56.406672+00:00 · methodology

0 comments
Cite this review

Pith. "Pith review of A Large-Language-Model Supported Personalized Driving Framework for Lane Change in Highway Scenarios." pith.science (2026). https://pith.science/paper/V7MJJFE6

@misc{pith2026260631483,
  author       = {Pith},
  title        = {Pith review of: A Large-Language-Model Supported Personalized Driving Framework for Lane Change in Highway Scenarios},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/V7MJJFE6}},
  note         = {Machine review of arXiv:2606.31483}
}
Share X Bluesky LinkedIn Reddit HN
read the original abstract

Personalized driving can improve the user acceptance of automated driving systems. However, existing methods still provide limited support for translating natural-language driving preferences, especially when such preferences are expressed implicitly, into executable and distinguishable driving behaviors. This paper proposes a large language model (LLM)-supported personalized driving framework for highway lane-change scenarios. The framework maps natural-language driving commands to executable planning parameters in the open-source Apollo automated driving stack according to three driving styles: aggressive, normal, and conservative. To establish this mapping, candidate planning parameters are evaluated based on the resulting lane-change behaviors, and style-specific parameter sets are constructed through clustering and style-intensity ranking. For command interpretation, a retrieval dataset is constructed to support retrieval-augmented generation (RAG), enabling LLM-based interpretation of implicit user commands. Experimental results show that the derived parameter sets generate distinguishable personalized lane-change behaviors, while RAG consistently improves preference interpretation, particularly for implicit commands. These results indicate the potential of integrating LLM-based natural-language interaction with Apollo to support personalized lane-change behavior generation. The source code and the relevant datasets are available at: https://github.com/ftgTUGraz/LLM-Personalized-Driving.

Figures

Figures reproduced from arXiv: 2606.31483 by Arno Eichberger, Dong Bi, Jiayuan Gong, Ji Zhou, Paul Kovacevic, Tomislav Mihalj, Yongqi Zhao.

Figure 1
Figure 1. Figure 1: Architecture of the proposed personalized driving simulation framework. [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: Workflow of LLM-based interaction program for driving style [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: Multi-agent data generation framework for constructing the RAG dataset used in driving style classification. [PITH_FULL_IMAGE:figures/full_fig_p005_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: A typical lane change scenario. 126 km/h, while all surrounding vehicles travel at 80 km/h. The initial longitudinal spacing is determined according to time-to-collision (TTC) criterion. A target TTC of 5.5 s is adopted, which is slightly larger than the 4.5–5 s range suggested for motorway collision-avoidance warning strategies [41]. The spacing is calculated as d = vrel · T T C. (8) where vrel denotes th… view at source ↗
Figure 5
Figure 5. Figure 5: Statistical distributions of lane-change time and lateral dynamic [PITH_FULL_IMAGE:figures/full_fig_p008_5.png] view at source ↗
Figure 7
Figure 7. Figure 7: Lane-change KPI distributions across driving styles. [PITH_FULL_IMAGE:figures/full_fig_p009_7.png] view at source ↗
Figure 6
Figure 6. Figure 6: Road network used for the experiments [PITH_FULL_IMAGE:figures/full_fig_p009_6.png] view at source ↗
Figure 8
Figure 8. Figure 8: The longitudinal speed planning process in Apollo. [PITH_FULL_IMAGE:figures/full_fig_p010_8.png] view at source ↗
Figure 9
Figure 9. Figure 9: The lane-change trajectory generation process in Apollo. [PITH_FULL_IMAGE:figures/full_fig_p011_9.png] view at source ↗

Review history (2 revisions) →

discussion (0)

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

Reference graph

Works this paper leans on

48 extracted references · 48 canonical work pages

  1. [2]

    A survey of personal- ization for advanced driver assistance systems,

    M. Hasenj ¨ager, M. Heckmann, and H. Wersing, “A survey of personal- ization for advanced driver assistance systems,”IEEE Transactions on Intelligent Vehicles, vol. 5, no. 2, pp. 335–344, 2020

    work page 2020

  2. [3]

    Implicit personalization in driving assistance: State-of-the-art and open issues,

    D. Yi, J. Su, L. Hu, C. Liu, M. Quddus, M. Dianati, and W.-H. Chen, “Implicit personalization in driving assistance: State-of-the-art and open issues,”IEEE Transactions on Intelligent Vehicles, vol. 5, no. 3, pp. 397–413, 2019

    work page 2019

  3. [4]

    A review of personalization in driving behavior: Dataset, modeling, and validation,

    X. Liao, Z. Zhao, M. J. Barth, A. Abdelraouf, R. Gupta, K. Han, J. Ma, and G. Wu, “A review of personalization in driving behavior: Dataset, modeling, and validation,”IEEE Transactions on Intelligent Vehicles, 2024

    work page 2024

  4. [5]

    Lane change and merge maneuvers for connected and automated vehicles: A survey,

    D. Bevly, X. Cao, M. Gordon, G. Ozbilgin, D. Kari, B. Nelson, J. Woodruff, M. Barth, C. Murray, A. Kurtet al., “Lane change and merge maneuvers for connected and automated vehicles: A survey,” IEEE Transactions on Intelligent Vehicles, vol. 1, no. 1, pp. 105–120, 2016

    work page 2016

  5. [6]

    Maveric: A data-driven approach to personalized autonomous driving,

    M. L. Schrum, E. Sumner, M. C. Gombolay, and A. Best, “Maveric: A data-driven approach to personalized autonomous driving,”IEEE Transactions on Robotics, vol. 40, pp. 1952–1965, 2024

    work page 1952

  6. [7]

    Personalized driver/vehicle lane change models for ADAS,

    V . A. Butakov and P. Ioannou, “Personalized driver/vehicle lane change models for ADAS,”IEEE Transactions on Vehicular Technology, vol. 64, no. 10, pp. 4422–4431, 2014

    work page 2014

  7. [8]

    A personalized human- like lane-changing trajectory planning method for automated driving system,

    S. Yang, H. Zheng, J. Wang, and A. El Kamel, “A personalized human- like lane-changing trajectory planning method for automated driving system,”IEEE Transactions on Vehicular Technology, vol. 70, no. 7, pp. 6399–6414, 2021

    work page 2021

  8. [9]

    Personalized trajectory planning and control of lane-change maneuvers for autonomous driving,

    C. Huang, H. Huang, P. Hang, H. Gao, J. Wu, Z. Huang, and C. Lv, “Personalized trajectory planning and control of lane-change maneuvers for autonomous driving,”IEEE Transactions on Vehicular Technology, vol. 70, no. 6, pp. 5511–5523, 2021. JOURNAL OF LATEX CLASS FILES, VOL. 14, NO. 8, AUGUST 2021 12

    work page 2021

  9. [10]

    Learning autonomous driving tasks via human feedbacks with large language models,

    Y . Ma, X. Cao, W. Ye, C. Cui, K. Mei, and Z. Wang, “Learning autonomous driving tasks via human feedbacks with large language models,” inFindings of the Association for Computational Linguistics: EMNLP 2024, 2024, pp. 4985–4995

    work page 2024

  10. [11]

    Autoware. flex: Human-instructed dynamically reconfigurable autonomous driving systems,

    Z. Song, M. Lv, T. Ren, C. J. Xue, J.-M. Wu, and N. Guan, “Autoware. flex: Human-instructed dynamically reconfigurable autonomous driving systems,” in2025 IEEE 31st International Conference on Embedded and Real-Time Computing Systems and Applications (RTCSA). IEEE, 2025, pp. 1–11

    work page 2025

  11. [12]

    Drive as you speak: Enabling human-like interaction with large language models in au- tonomous vehicles,

    C. Cui, Y . Ma, X. Cao, W. Ye, and Z. Wang, “Drive as you speak: Enabling human-like interaction with large language models in au- tonomous vehicles,” inProceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision, 2024, pp. 902–909

    work page 2024

  12. [13]

    Dilu: A knowledge-driven approach to autonomous driving with large language models,

    L. Wen, D. Fu, X. Li, X. Cai, T. Ma, P. Cai, M. Dou, B. Shi, L. He, and Y . Qiao, “Dilu: A knowledge-driven approach to autonomous driving with large language models,” inInternational Conference on Learning Representations, 2024, pp. 34 503–34 522

    work page 2024

  13. [14]

    StyleVLA: Driving style-aware vision language action model for autonomous driving.arXiv preprint arXiv:2603.09482, 2026

    Y . Gao, D. Hua, M. Piccinini, F. R. Sch ¨afer, K. Moller, L. Li, and J. Betz, “Stylevla: Driving style-aware vision language action model for autonomous driving,”arXiv preprint arXiv:2603.09482, 2026

    work page arXiv 2026

  14. [15]

    Padriver: Towards personalized autonomous driving,

    G. Kou, F. Jia, W. Mao, Y . Liu, Y . Zhao, Z. Zhang, O. Yoshie, T. Wang, Y . Li, and X. Zhang, “Padriver: Towards personalized autonomous driving,” in2025 International Joint Conference on Neural Networks (IJCNN). IEEE, 2025, pp. 1–8

    work page 2025

  15. [16]

    A survey of personal- ization for advanced driver assistance systems,

    M. Hasenj ¨ager, M. Heckmann, and H. Wersing, “A survey of personal- ization for advanced driver assistance systems,”IEEE Transactions on Intelligent Vehicles, vol. 5, no. 2, pp. 335–344, 2019

    work page 2019

  16. [17]

    Apollo: An open autonomous driving platform,

    ApolloAuto, “Apollo: An open autonomous driving platform,” https: //github.com/ApolloAuto/apollo, gitHub repository. Accessed: Jun. 21, 2026

    work page 2026

  17. [18]

    Personalized forward collision warning model with learning from human preferences,

    N. Xie, R. Yu, W. Sun, S. Qiu, K. Zhong, M. Xu, G. Wu, and Y . Yang, “Personalized forward collision warning model with learning from human preferences,”Accident Analysis & Prevention, vol. 208, p. 107791, 2024

    work page 2024

  18. [19]

    Personalized adaptive cruise control via gaussian process regression,

    Y . Wang, Z. Wang, K. Han, P. Tiwari, and D. B. Work, “Personalized adaptive cruise control via gaussian process regression,” in2021 IEEE International Intelligent Transportation Systems Conference (ITSC). IEEE, 2021, pp. 1496–1502

    work page 2021

  19. [20]

    Drivers’ inter- action with adaptive cruise control on dry and snowy roads with various tire-road grip potentials,

    I. Koglbauer, J. Holzinger, A. Eichberger, and C. Lex, “Drivers’ inter- action with adaptive cruise control on dry and snowy roads with various tire-road grip potentials,”Journal of Advanced Transportation, vol. 2017, no. 1, p. 5496837, 2017

    work page 2017

  20. [21]

    Personalizing automated driv- ing speed to enhance user experience and performance in intermediate- level automated driving,

    M. Delmas, V . Camps, and C. Lemercier, “Personalizing automated driv- ing speed to enhance user experience and performance in intermediate- level automated driving,”Accident Analysis & Prevention, vol. 199, p. 107512, 2024

    work page 2024

  21. [22]

    User-driven adaptation: Tailoring autonomous driving systems with dynamic preferences,

    M. Zhang, J. Li, N. Li, E. Kang, and K. Tei, “User-driven adaptation: Tailoring autonomous driving systems with dynamic preferences,” in Extended Abstracts of the CHI Conference on Human Factors in Computing Systems, 2024, pp. 1–8

    work page 2024

  22. [23]

    Accelerating the evolution of personalized automated lane change through lesson learning,

    J. Hu, M. Lei, H. Wang, Z. Liu, and F. Yang, “Accelerating the evolution of personalized automated lane change through lesson learning,”IEEE Transactions on Intelligent Transportation Systems, 2025

    work page 2025

  23. [24]

    Large language models for autonomous driving—concept, review, benchmark, experiments, and future trends,

    C. Cui, Y . Ma, S.-Y . Park, Z. Yang, Y . Zhou, P. Liu, J. Lu, J. Peng, J. Zhang, R. Zhanget al., “Large language models for autonomous driving—concept, review, benchmark, experiments, and future trends,” Proceedings of the IEEE, 2026

    work page 2026

  24. [25]

    Llm4drive: A survey of large language models for autonomous driving.arXiv preprint arXiv:2311.01043, 2023

    Z. Yang, X. Jia, H. Li, and J. Yan, “Llm4drive: A survey of large language models for autonomous driving,”arXiv preprint arXiv:2311.01043, 2023

    work page arXiv 2023

  25. [26]

    Vision language models in autonomous driving: A survey and outlook,

    X. Zhou, M. Liu, E. Yurtsever, B. L. Zagar, W. Zimmer, H. Cao, and A. C. Knoll, “Vision language models in autonomous driving: A survey and outlook,”IEEE Transactions on Intelligent Vehicles, 2024

    work page 2024

  26. [27]

    A survey on the application of large language models in scenario-based testing of automated driving systems,

    Y . Zhao, J. Zhou, D. Bi, T. Mihalj, J. Hu, and A. Eichberger, “A survey on the application of large language models in scenario-based testing of automated driving systems,”IEEE Transactions on Intelligent Transportation Systems, 2026

    work page 2026

  27. [28]

    Personalized autonomous driving with large language models: Field experiments,

    C. Cui, Z. Yang, Y . Zhou, Y . Ma, J. Lu, L. Li, Y . Chen, J. Panchal, and Z. Wang, “Personalized autonomous driving with large language models: Field experiments,” in2024 IEEE 27th International Conference on Intelligent Transportation Systems (ITSC). IEEE, 2024, pp. 20–27

    work page 2024

  28. [29]

    Personalizing driver agent using large language models for driving safety and smarter human–machine interactions,

    Z. Xu, T. Chen, Z. Huang, Y . Xing, and S. Chen, “Personalizing driver agent using large language models for driving safety and smarter human–machine interactions,”IEEE intelligent transportation Systems magazine, 2025

    work page 2025

  29. [30]

    Carla: An open urban driving simulator,

    A. Dosovitskiy, G. Ros, F. Codevilla, A. Lopez, and V . Koltun, “Carla: An open urban driving simulator,” inConference on robot learning. PMLR, 2017, pp. 1–16

    work page 2017

  30. [31]

    Llm-based operating systems for automated vehicles: A new perspective,

    J. Ge, C. Chang, J. Zhang, L. Li, X. Na, Y . Lin, L. Li, and F.-Y . Wang, “Llm-based operating systems for automated vehicles: A new perspective,”IEEE Transactions on Intelligent Vehicles, vol. 9, no. 4, pp. 4563–4567, 2024

    work page 2024

  31. [32]

    Human-centric autonomous systems with llms for user command reasoning,

    Y . Yang, Q. Zhang, C. Li, D. S. Marta, N. Batool, and J. Folkesson, “Human-centric autonomous systems with llms for user command reasoning,” inProceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision, 2024, pp. 988–994

    work page 2024

  32. [33]

    Gpt- 4 enhanced multimodal grounding for autonomous driving: Leveraging cross-modal attention with large language models,

    H. Liao, H. Shen, Z. Li, C. Wang, G. Li, Y . Bie, and C. Xu, “Gpt- 4 enhanced multimodal grounding for autonomous driving: Leveraging cross-modal attention with large language models,”Communications in Transportation Research, vol. 4, p. 100116, 2024

    work page 2024

  33. [34]

    doscenes: An autonomous driving dataset with natural lan- guage instruction for human interaction and vision-language navigation,

    P. Roy, S. Perisetla, S. Shriram, H. Krishnaswamy, A. Keskar, and R. Greer, “doscenes: An autonomous driving dataset with natural lan- guage instruction for human interaction and vision-language navigation,” in2025 IEEE 28th International Conference on Intelligent Transporta- tion Systems (ITSC). IEEE, 2025, pp. 1651–1658

    work page 2025

  34. [35]

    Talk2car: Taking control of your self-driving car,

    T. Deruyttere, S. Vandenhende, D. Grujicic, L. Van Gool, and M. F. Moens, “Talk2car: Taking control of your self-driving car,” inPro- ceedings of the 2019 conference on empirical methods in natural language processing and the 9th international joint conference on natural language processing (EMNLP-IJCNLP), 2019, pp. 2088–2098

    work page 2019

  35. [36]

    Language prompt for autonomous driving,

    D. Wu, W. Han, Y . Liu, T. Wang, C.-z. Xu, X. Zhang, and J. Shen, “Language prompt for autonomous driving,” inProceedings of the AAAI conference on artificial intelligence, vol. 39, no. 8, 2025, pp. 8359–8367

    work page 2025

  36. [37]

    Lmdrive: Closed-loop end-to-end driving with large language models,

    H. Shao, Y . Hu, L. Wang, G. Song, S. L. Waslander, Y . Liu, and H. Li, “Lmdrive: Closed-loop end-to-end driving with large language models,” inProceedings of the IEEE/CVF conference on computer vision and pattern recognition, 2024, pp. 15 120–15 130

    work page 2024

  37. [38]

    sentence-transformers/all-MiniLM-L6-v2,

    Sentence Transformers, “sentence-transformers/all-MiniLM-L6-v2,” [Online]. Available: https://huggingface.co/sentence-transformers/ all-MiniLM-L6-v2, 2021, accessed: Jun. 15, 2026

    work page 2021

  38. [39]

    The faiss library,

    M. Douze, A. Guzhva, C. Deng, J. Johnson, G. Szilvasy, P.-E. Mazar ´e, M. Lomeli, L. Hosseini, and H. J ´egou, “The faiss library,” 2024

    work page 2024

  39. [40]

    Billion-scale similarity search with GPUs,

    J. Johnson, M. Douze, and H. J ´egou, “Billion-scale similarity search with GPUs,”IEEE Transactions on Big Data, vol. 7, no. 3, pp. 535– 547, 2019

    work page 2019

  40. [41]

    CarMaker: The simulation solution for virtual test driving,

    IPG Automotive GmbH, “CarMaker: The simulation solution for virtual test driving,” https://www.ipg-automotive.com/solutions/ product-portfolio/carmaker, accessed: Jun. 21, 2026

    work page 2026

  41. [42]

    Toward a full-stack co-simulation platform for testing of automated driving systems,

    D. Bi, Y . Zhao, Z. Gu, T. Mihalj, J. Hu, and A. Eichberger, “Toward a full-stack co-simulation platform for testing of automated driving systems,” in2025 IEEE 28th International Conference on Intelligent Transportation Systems (ITSC), 2025, pp. 980–986

    work page 2025

  42. [43]

    Chat2scenario: Scenario extraction from dataset through utilization of large language model,

    Y . Zhao, W. Xiao, T. Mihalj, J. Hu, and A. Eichberger, “Chat2scenario: Scenario extraction from dataset through utilization of large language model,” in2024 IEEE Intelligent Vehicles Symposium (IV), 2024, pp. 559–566

    work page 2024

  43. [44]

    Research on car-following model considering driving style,

    K. Wang, Y . Yang, S. Wang, and Z. Shi, “Research on car-following model considering driving style,”Mathematical Problems in Engineer- ing, vol. 2022, no. 1, p. 7215697, 2022

    work page 2022

  44. [45]

    Optimal trajectory generation for dynamic street scenarios in a frenet frame,

    M. Werling, J. Ziegler, S. Kammel, and S. Thrun, “Optimal trajectory generation for dynamic street scenarios in a frenet frame,” in2010 IEEE international conference on robotics and automation. IEEE, 2010, pp. 987–993

    work page 2010

  45. [46]

    Personalized lane-change as- sistance system with driver behavior identification,

    B. Zhu, S. Yan, J. Zhao, and W. Deng, “Personalized lane-change as- sistance system with driver behavior identification,”IEEE Transactions on Vehicular Technology, vol. 67, no. 11, pp. 10 293–10 306, 2018

    work page 2018

  46. [47]

    Some methods of classification and analysis of mul- tivariate observations,

    J. B. McQueen, “Some methods of classification and analysis of mul- tivariate observations,” inProc. of 5th Berkeley Symposium on Math. Stat. and Prob., 1967, pp. 281–297

    work page 1967

  47. [48]

    Development of a co-simulation framework for systematic generation of scenarios for testing and validation of automated driving systems,

    D. Nalic, A. Eichberger, G. Hanzl, M. Fellendorf, and B. Rogic, “Development of a co-simulation framework for systematic generation of scenarios for testing and validation of automated driving systems,” in 2019 IEEE Intelligent Transportation Systems Conference (ITSC), 2019, pp. 1895–1901

    work page 2019

  48. [49]

    A communication-latency-aware co-simulation platform for safety and comfort evaluation of cloud-controlled icvs,

    Y . Zhao, X. Zhang, T. Mihalj, M. Schabauer, L. Putzer, E. Reichmann- Blaga, ´A. Borony´ak, A. R ¨ovid, G. So ´os, P. Zhang, L. Xiong, J. Hu, and A. Eichberger, “A communication-latency-aware co-simulation platform for safety and comfort evaluation of cloud-controlled icvs,”IEEE Internet of Things Journal, vol. 13, no. 4, pp. 6217–6229, 2026. JOURNAL OF L...

    work page 2026