pith. sign in

arxiv: 2605.12321 · v2 · pith:Q55FZGYVnew · submitted 2026-05-12 · 💻 cs.AI · cs.CY· cs.ET

LIDSA: Cognitive Arbitration for Signal-Free Autonomous Intersection Management

Abderrahmane Lakas , Mohamed Amine Ferrag , Merouane Debbah This is my paper

Pith reviewed 2026-05-21 08:14 UTC · model grok-4.3

classification 💻 cs.AI cs.CYcs.ET
keywords autonomous intersection managementlarge language modelssignal-free controlintent-driven speed advisorytraffic simulationcognitive arbitrationmulti-agent coordination
0
0 comments X

The pith

An LLM can directly arbitrate right-of-way at intersections without traffic lights by reasoning over vehicle intents.

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

The paper introduces LIDSA as a signal-free framework that places a large language model at the center of intersection decisions. The model receives declared intents from approaching vehicles and weighs factors like priority class, current queue pressure, and energy preferences to issue speed advisories that resolve conflicts in real time. Simulations compare it to fixed-cycle signals, SCATS, AIM, and GLOSA across different demand levels, showing that the LLM approach sustains acceptable flow while the others collapse. If the claim holds, language-model reasoning could replace rigid, vehicle-agnostic controllers in settings where rules alone cannot adapt to heterogeneous vehicle goals and kinematics.

Core claim

LIDSA uses an LLM to reason over declared vehicle intents, incorporating priority classes, queue pressure, and energy preferences, to generate speed advisories that enable real-time, signal-free intersection management, achieving up to 89.1 percent lower mean control delay, Level of Service C under loads that degrade baselines to Level of Service F, 93 percent shorter mean waiting time, 60.6 percent smaller peak queues, 48.8 percent lower fuel use, and 86.2 percent intent satisfaction compared with the best non-LLM method.

What carries the argument

LIDSA framework, in which an LLM performs cognitive arbitration by processing vehicle intents and producing speed advisories without any signal infrastructure.

If this is right

  • Intersections can sustain Level of Service C even near capacity instead of degrading to F.
  • Mean vehicle waiting time drops by more than 90 percent relative to fixed-cycle timing.
  • Peak queue lengths shrink by over 60 percent, reducing spillback onto upstream links.
  • Fuel consumption falls by nearly half through smoother speed profiles.

Where Pith is reading between the lines

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

  • The same intent-reasoning pattern could be tested at coordinated corridor level rather than isolated intersections.
  • Mixed fleets containing human-driven vehicles would require a fallback mechanism when intents cannot be declared reliably.
  • Hardware-in-the-loop experiments with actual vehicle kinematics would reveal whether the reported fuel and delay gains survive communication jitter.

Load-bearing premise

LLM inference can be performed fast enough for sub-second control decisions and that simulated vehicle intents match real-world behavior without communication failures.

What would settle it

A physical testbed measurement of end-to-end LLM response latency and safety incidents when vehicles follow the generated speed advisories at a real intersection under near-saturated flow.

Figures

Figures reproduced from arXiv: 2605.12321 by Abderrahmane Lakas, Merouane Debbah, Mohamed Amine Ferrag.

Figure 1
Figure 1. Figure 1: Free-signal Intersection Management using LLM-based [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: LISA architecture for signal-free, intent-based intersection management, combining LLM-driven arbitration, cache-based [PITH_FULL_IMAGE:figures/full_fig_p005_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: A signal-free four-way intersection with the advisory [PITH_FULL_IMAGE:figures/full_fig_p006_3.png] view at source ↗
Figure 5
Figure 5. Figure 5: Peak inbound queue length (veh). and SCATS degrade between Medium and High load because signalized operation becomes oversaturated. Mean control delay is the clearest efficiency differentiator. As shown in [PITH_FULL_IMAGE:figures/full_fig_p011_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: Overall intent satisfaction rate (%). The score combines [PITH_FULL_IMAGE:figures/full_fig_p012_6.png] view at source ↗
Figure 7
Figure 7. Figure 7: Average fuel consumed per vehicle (g/veh). [PITH_FULL_IMAGE:figures/full_fig_p013_7.png] view at source ↗
read the original abstract

Large language models (LLMs) show strong potential for Intelligent Transportation Systems (ITS), particularly in tasks requiring situational reasoning and multi-agent coordination. These capabilities make them well suited for cooperative driving, where rule-based approaches struggle in complex and dynamic traffic environments. Intersection management remains especially challenging due to conflicting right-of-way demands, heterogeneous vehicle priorities, and vehicle-specific kinematic constraints that must be resolved in real time. However, existing approaches typically use LLMs as auxiliary components on top of signal-based systems rather than as primary decision-makers. Signal controllers remain vehicle-agnostic, reservation-based methods lack intent awareness, and recent LLM-based systems still depend on signal infrastructure. In addition, LLM inference latency limits their use in sub-second control settings. We propose LIDSA (LLM-Based Intent-Driven Speed Advisory), a signal-free cognitive arbitration framework for autonomous intersection management. LIDSA uses an LLM to reason over declared vehicle intents, incorporating priority classes, queue pressure, and energy preferences. We evaluate LIDSA against fixed-cycle control, SCATS, AIM, and GLOSA across varying traffic loads. Results show that LIDSA reduces mean control delay by up to 89.1% and maintains Level of Service C while all non-LLM baselines degrade to Level of Service F. Under near-saturated demand, LIDSA reduces mean waiting time by 93% and peak queue length by 60.6% relative to fixed-cycle control. It also lowers fuel consumption by up to 48.8% and achieves 86.2% intent satisfaction, compared to 61.2% for the best non-LLM method. These results demonstrate that LLM-based reasoning can enable real-time, signal-free intersection management.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

2 major / 1 minor

Summary. The manuscript proposes LIDSA (LLM-Based Intent-Driven Speed Advisory), a signal-free cognitive arbitration framework for autonomous intersection management. An LLM reasons over declared vehicle intents, priority classes, queue pressure, and energy preferences to generate speed advisories. The approach is evaluated in simulation against fixed-cycle control, SCATS, AIM, and GLOSA baselines across traffic loads, with claims of up to 89.1% reduction in mean control delay (maintaining LOS C while baselines reach LOS F), 93% reduction in mean waiting time, 60.6% reduction in peak queue length, up to 48.8% lower fuel consumption, and 86.2% intent satisfaction under near-saturated demand.

Significance. If the simulation results are reproducible and the real-time latency constraint is demonstrably satisfied, the work would provide evidence that LLM-based situational reasoning can serve as a primary controller for signal-free intersections, offering substantial gains over both traditional signal-based and reservation-based methods in delay, queueing, and intent compliance. The quantitative improvements are large enough to warrant further investigation if the evaluation methodology is clarified.

major comments (2)
  1. [Abstract] Abstract: The performance claims (89.1% delay reduction, LOS C maintenance, 93% waiting-time reduction) are presented from simulation without any reported information on the simulation platform, traffic demand model validation, number of runs, statistical tests, or how LLM outputs are mapped to vehicle speed advisories within the control loop. These details are required to assess whether the results support the real-time superiority claim.
  2. [Abstract] Abstract: The central claim requires sub-second real-time decisions, yet the manuscript acknowledges that LLM inference latency limits use in sub-second settings and provides no measurements of end-to-end decision latency, prompt construction time, or arbitration overhead. Without these data, it is not possible to confirm that the reported gains are achievable within the simulation timestep or in a deployable system.
minor comments (1)
  1. [Abstract] The abstract would benefit from a concise statement of the specific LLM employed and any prompt-engineering or output-parsing techniques used to ensure deterministic speed-advisory generation.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the constructive and detailed feedback. We have revised the manuscript to improve the clarity of the evaluation methodology and to provide explicit measurements addressing real-time feasibility. Our point-by-point responses follow.

read point-by-point responses

  1. Referee: [Abstract] Abstract: The performance claims (89.1% delay reduction, LOS C maintenance, 93% waiting-time reduction) are presented from simulation without any reported information on the simulation platform, traffic demand model validation, number of runs, statistical tests, or how LLM outputs are mapped to vehicle speed advisories within the control loop. These details are required to assess whether the results support the real-time superiority claim.

    Authors: We agree that the abstract would benefit from greater methodological transparency. The full manuscript already details the simulation platform, demand model validation against empirical data, use of multiple independent runs with statistical testing, and the precise interface mapping LLM outputs to speed advisories (see Sections 4 and 5). To make these elements immediately accessible from the abstract, we have added a concise summary of the evaluation protocol and control-loop integration in the revised abstract. revision: yes

  2. Referee: [Abstract] Abstract: The central claim requires sub-second real-time decisions, yet the manuscript acknowledges that LLM inference latency limits use in sub-second settings and provides no measurements of end-to-end decision latency, prompt construction time, or arbitration overhead. Without these data, it is not possible to confirm that the reported gains are achievable within the simulation timestep or in a deployable system.

    Authors: We accept this criticism. The original manuscript notes the general latency constraint but does not report concrete measurements. In the revision we have added a new latency analysis subsection that quantifies end-to-end decision time (inference, prompt assembly, and arbitration overhead) under the exact simulation conditions used for the performance results. These measurements confirm compatibility with the control timestep employed, and we have updated both the abstract and the discussion to clarify the intended operating regime and scope for real-world deployment. revision: yes

Circularity Check

0 steps flagged

No significant circularity in derivation or evaluation chain

full rationale

The paper introduces the LIDSA framework as a signal-free intersection management approach using LLM-based reasoning over vehicle intents and evaluates it via direct simulation comparisons to independent baselines (fixed-cycle control, SCATS, AIM, GLOSA). Reported gains such as 89.1% delay reduction and 93% waiting time reduction are obtained from these external benchmark runs rather than any fitted parameters, self-referential equations, or self-citation chains that reduce the outputs to the inputs by construction. No mathematical derivations, uniqueness theorems, or ansatzes are presented that could create self-definitional loops; the central claims rest on observable simulation outcomes against non-LLM methods.

Axiom & Free-Parameter Ledger

0 free parameters · 2 axioms · 0 invented entities

The central performance claims rest on the assumption that the simulation environment and LLM prompting strategy faithfully capture real-time traffic dynamics; no explicit free parameters or invented entities are stated in the abstract.

axioms (2)
  • domain assumption LLM inference can be completed within the time window required for sub-second vehicle control decisions
    Abstract notes that inference latency limits use in sub-second settings yet still claims real-time operation.
  • domain assumption Declared vehicle intents are truthful and accurately reflect kinematic constraints and priorities
    Framework relies on vehicles declaring intents that the LLM then arbitrates.

pith-pipeline@v0.9.0 · 5852 in / 1326 out tokens · 29888 ms · 2026-05-21T08:14:22.420494+00:00 · methodology

Review history (2 revisions) →

discussion (0)

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

Lean theorems connected to this paper

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

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

Reference graph

Works this paper leans on

42 extracted references · 42 canonical work pages · 3 internal anchors

  1. [1]

    Computational intelligence in urban traffic signal control: A survey,

    D. Zhao, Y . Dai, and Z. Zhang, “Computational intelligence in urban traffic signal control: A survey,”Trans. Sys. Man Cyber Part C, vol. 42, no. 4, pp. 485–494, Jul. 2012. [Online]. Available: https://doi.org/10.1109/TSMCC.2011.2161577

    work page doi:10.1109/tsmcc.2011.2161577 2012

  2. [2]

    Parallel transportation systems: Toward iot-enabled smart urban traffic control and management,

    F. Zhu, Y . Lv, Y . Chen, X. Wang, G. Xiong, and F.-Y . Wang, “Parallel transportation systems: Toward iot-enabled smart urban traffic control and management,”IEEE Transactions on Intelligent Transportation Systems, vol. 21, no. 10, pp. 4063–4071, Oct 2020

    work page 2020

  3. [3]

    A survey of autonomous vehicles: Enabling communication technologies and challenges,

    M. N. Ahangar, Q. Z. Ahmed, F. A. Khan, and M. Hafeez, “A survey of autonomous vehicles: Enabling communication technologies and challenges,”Sensors, vol. 21, no. 3, 2021. [Online]. Available: https://www.mdpi.com/1424-8220/21/3/706

    work page 2021

  4. [4]

    Cooperative intersection management: A sur- vey,

    L. Chen and C. Englund, “Cooperative intersection management: A sur- vey,”IEEE Transactions on Intelligent Transportation Systems, vol. 17, no. 2, pp. 570–586, Feb 2016

    work page 2016

  5. [5]

    Intelligent intersection management sys- tems considering autonomous vehicles: A systematic literature review,

    E. Namazi, J. Li, and C. Lu, “Intelligent intersection management sys- tems considering autonomous vehicles: A systematic literature review,” IEEE Access, vol. 7, pp. 91 946–91 965, 2019

    work page 2019

  6. [6]

    Queue-responsive adaptive signal con- trol vs. Webster optimization: A multi-criteria simulation assessment at a signalized intersection,

    M. Albdairi and A. Almusawi, “Queue-responsive adaptive signal con- trol vs. Webster optimization: A multi-criteria simulation assessment at a signalized intersection,”Future Transportation, vol. 6, no. 2, p. 92, 2026

    work page 2026

  7. [7]

    Green-pressure – a weighted queue-length approach towards sustainable intersection man- agement,

    K. Riehl, A. Kouvelas, and M. A. Makridis, “Green-pressure – a weighted queue-length approach towards sustainable intersection man- agement,” in2025 IEEE 64th Conference on Decision and Control (CDC), Dec 2025, pp. 4964–4969

    work page 2025

  8. [8]

    The traffic signal control problem for inter- sections: a review,

    M. Eom and B.-I. Kim, “The traffic signal control problem for inter- sections: a review,”European transport research review, vol. 12, no. 1, p. 50, 2020

    work page 2020

  9. [9]

    State-of-art review of traffic signal control methods: challenges and opportunities,

    S. S. S. M. Qadri, M. A. G ¨okc ¸e, and E. ¨Oner, “State-of-art review of traffic signal control methods: challenges and opportunities,”European transport research review, vol. 12, no. 1, p. 55, 2020

    work page 2020

  10. [10]

    Green light optimized speed advisory systems: Impact of signal phasing information accuracy,

    A. Stevanovic, J. Stevanovic, and C. Kergaye, “Green light optimized speed advisory systems: Impact of signal phasing information accuracy,” Transportation research record, vol. 2390, no. 1, pp. 53–59, 2013

    work page 2013

  11. [11]

    Potentials and limitations of green light optimal speed advisory systems,

    D. Eckhoff, B. Halmos, and R. German, “Potentials and limitations of green light optimal speed advisory systems,” in2013 IEEE Vehicular Networking Conference. IEEE, 2013, pp. 103–110

    work page 2013

  12. [12]

    Assessing connected vehicle’s response to green light optimal speed advisory from field operational test and scaling up,

    K. Bhattacharyya, P.-A. Laharotte, A. Burianne, and N.-E. E. Faouzi, “Assessing connected vehicle’s response to green light optimal speed advisory from field operational test and scaling up,”IEEE Transactions on Intelligent Transportation Systems, vol. 24, no. 6, pp. 6725–6736, June 2023

    work page 2023

  13. [13]

    A multiagent approach to autonomous intersection management,

    K. Dresner and P. Stone, “A multiagent approach to autonomous intersection management,”Journal of Artificial Intelligence Research, vol. 31, pp. 591–656, 2008

    work page 2008

  14. [14]

    A survey on intersection management of connected autonomous vehicles,

    M. Khayatian, M. Mehrabian, E. Andert, R. Dedinsky, S. Choudhary, Y . Lou, and A. Shirvastava, “A survey on intersection management of connected autonomous vehicles,”ACM Transactions on Cyber-Physical Systems, vol. 4, no. 4, pp. 1–27, 2020

    work page 2020

  15. [15]

    A comprehensive survey on coop- erative intersection management for heterogeneous connected vehicles,

    A. Gholamhosseinian and J. Seitz, “A comprehensive survey on coop- erative intersection management for heterogeneous connected vehicles,” IEEE Access, vol. 10, pp. 7937–7972, 2022

    work page 2022

  16. [16]

    Dcl-aim: Decentralized coordination learning of autonomous intersection management for connected and automated vehicles,

    Y . Wu, H. Chen, and F. Zhu, “Dcl-aim: Decentralized coordination learning of autonomous intersection management for connected and automated vehicles,”Transportation Research Part C: Emerging Tech- nologies, vol. 103, pp. 246–260, 2019

    work page 2019

  17. [17]

    Llm and ai agents for autonomous systems: A survey of applications, datasets, and security challenges,

    M. A. Ferrag, A. Lakas, N. Tihanyi, and M. Debbah, “Llm and ai agents for autonomous systems: A survey of applications, datasets, and security challenges,”IEEE Open Journal of Intelligent Transportation Systems, vol. 7, pp. 615–657, 2026

    work page 2026

  18. [18]

    Large (vision) language models for autonomous vehicles: Current trends and future directions,

    H. Tian, K. Reddy, Y . Feng, M. Quddus, Y . Demiris, and P. An- geloudis, “Large (vision) language models for autonomous vehicles: Current trends and future directions,”IEEE Transactions on Intelligent Transportation Systems, vol. 27, no. 1, pp. 187–210, Jan 2026

    work page 2026

  19. [19]

    Multimodal large language models for perception in autonomous driving: Architecture, taxonomy, and challenges,

    Y . Jing, X. Zhang, M. Zhou, Y . Ji, Y . Ding, J. Yin, R. Jia, Y . Xiong, K. Zhao, J. Yang, J. Li, and H. Liu, “Multimodal large language models for perception in autonomous driving: Architecture, taxonomy, and challenges,”IEEE Internet of Things Journal, vol. 13, no. 10, pp. 20 375–20 397, May 2026

    work page 2026

  20. [20]

    The crossroads of LLM and traffic control: A study on large language models in adaptive traffic signal control,

    M. Movahedi and J. Choi, “The crossroads of LLM and traffic control: A study on large language models in adaptive traffic signal control,”IEEE Transactions on Intelligent Transportation Systems, vol. 26, no. 2, pp. 1701–1716, February 2025. [Online]. Available: https://ieeexplore.ieee.org/abstract/document/10768207/

    work page arXiv 2025

  21. [21]

    LLMLight: Large language models as traffic signal control agents,

    S. Lai, Z. Xu, W. Zhang, H. Liu, and H. Xiong, “LLMLight: Large language models as traffic signal control agents,” inProceedings of the 31st ACM SIGKDD Conference on Knowledge Discovery and Data Mining. ACM, 2025. [Online]. Available: https://dl.acm.org/doi/abs/ 10.1145/3690624.3709379

    work page doi:10.1145/3690624.3709379 2025

  22. [22]

    illm-tsc: Integration reinforcement learning and large language model for traffic signal control policy improvement,

    A. Pang, M. Wang, M.-O. Pun, C. S. Chen, and X. Xiong, “illm-tsc: Integration reinforcement learning and large language model for traffic signal control policy improvement,”IEEE Transactions on Vehicular Technology, pp. 1–14, 2026

    work page 2026

  23. [23]

    A large language model agent-guided multi-agent system for adaptive traffic signal control,

    M. Zhu and C. Zhu, “A large language model agent-guided multi-agent system for adaptive traffic signal control,” inKnowledge Science, Engineering and Management (KSEM 2025), ser. Lecture Notes in Computer Science (LNAI), vol. 15919. Springer, November 2025, pp. 59–70. [Online]. Available: https://link.springer.com/chapter/10.1007/ 978-981-95-3001-4\ 5

    work page 2025

  24. [24]

    Large language models (LLMs) as traffic control systems at urban intersections: A new paradigm,

    S. Masri, H. I. Ashqar, and M. Elhenawy, “Large language models (LLMs) as traffic control systems at urban intersections: A new paradigm,”Vehicles, vol. 7, no. 1, p. 11, 2025. [Online]. Available: https://www.mdpi.com/2624-8921/7/1/11

    work page 2025

  25. [25]

    Enhancing traffic sustainability: An analysis of isolation intersection effectiveness through fixed time and logic control design using visvap algorithm,

    R. Duraku and D. Boshnjaku, “Enhancing traffic sustainability: An analysis of isolation intersection effectiveness through fixed time and logic control design using visvap algorithm,”Sustainability, vol. 16, no. 7, p. 2930, 2024

    work page 2024

  26. [26]

    Adaptive traffic signal timing optimization using deep reinforcement learning in urban networks,

    Z. Wu, S. Wang, C. Ni, and J. Wu, “Adaptive traffic signal timing optimization using deep reinforcement learning in urban networks,” Artificial Intelligence and Machine Learning Research, 2024. [Online]. Available: https://scipublication.com/index.php/AIMLR/article/view/119

    work page 2024

  27. [27]

    Decentralized cycle-free game-theoretic adaptive traffic signal control: Model enhancement and testing on isolated signalized intersections,

    A. K. Shafik and H. A. Rakha, “Decentralized cycle-free game-theoretic adaptive traffic signal control: Model enhancement and testing on isolated signalized intersections,”Sensors, vol. 25, no. 20, p. 6339, 2025

    work page 2025

  28. [28]

    Adaptive traffic signal control using genetic algorithm for a 2×2 traffic network,

    M. K. Tan, S. Q. Chai, H. S. Ee Chuo, K. G. Lim, H. H. Goh, and K. T. K. Teo, “Adaptive traffic signal control using genetic algorithm for a 2×2 traffic network,” in2024 IEEE International Conference on Artificial Intelligence in Engineering and Technology (IICAIET), Aug 2024, pp. 488–493

    work page 2024

  29. [29]

    A decentralized signal-free intersection control framework for connected and autonomous vehicles,

    H. Guan, Q. Bai, and Q. Meng, “A decentralized signal-free intersection control framework for connected and autonomous vehicles,”IEEE Transactions on Consumer Electronics, vol. 71, no. 2, pp. 6193–6200, May 2025

    work page 2025

  30. [30]

    Centrally coordinated vehicles in a signal free intersection,

    G. Rostomyan, K. Savla, and P. A. Ioannou, “Centrally coordinated vehicles in a signal free intersection,” in2024 IEEE 27th International Conference on Intelligent Transportation Systems (ITSC), Sep. 2024, pp. 3279–3284

    work page 2024

  31. [31]

    Cooperative driving at multiple unsignalized intersections in fully autonomous driv- ing scenarios,

    W. Pan, B. Lin, Y . Wang, Z. Yu, X. Zhao, X. He, and J. Ye, “Cooperative driving at multiple unsignalized intersections in fully autonomous driv- ing scenarios,”IEEE Transactions on Intelligent Transportation Systems, 2025

    work page 2025

  32. [32]

    Artificial intelligence-based adaptive traffic signal control system: A comprehensive review,

    A. Agrahari, M. M. Dhabu, P. S. Deshpande, A. Tiwari, M. A. Baig, and A. D. Sawarkar, “Artificial intelligence-based adaptive traffic signal control system: A comprehensive review,”Electronics, vol. 13, no. 19, p. 3875, 2024

    work page 2024

  33. [33]

    A compre- hensive survey of adaptive traffic signal control: Methods, applications, challenges, and future research,

    M. Hassan, T. D. Turjo, A. H. Rambe, M. Jusoh, M. K. Islam, M. Hafizur Rahman, M. S. Alam, and M. M. Rahman, “A compre- hensive survey of adaptive traffic signal control: Methods, applications, challenges, and future research,”Archives of Computational Methods in Engineering, pp. 1–64, 2026

    work page 2026

  34. [34]

    On 5g-v2x use cases and enabling technologies: A comprehensive survey,

    A. Alalewi, I. Dayoub, and S. Cherkaoui, “On 5g-v2x use cases and enabling technologies: A comprehensive survey,”IEEE Access, vol. 9, pp. 107 710–107 737, 2021

    work page 2021

  35. [35]

    Influence of cav clustering strategies on mixed traffic flow characteristics: An analysis of vehicle trajectory data,

    Z. Zhong, E. E. Lee, M. Nejad, and J. Lee, “Influence of cav clustering strategies on mixed traffic flow characteristics: An analysis of vehicle trajectory data,”Transportation Research Part C: Emerging Technologies, vol. 115, p. 102611, 2020. [Online]. Available: https://www.sciencedirect.com/science/article/pii/S0968090X19307648

    work page 2020

  36. [36]

    Measuring massive multitask language understanding,

    D. Hendrycks, C. Burns, S. Basart, A. Zou, M. Mazeika, D. Song, and J. Steinhardt, “Measuring massive multitask language understanding,”

    work page

  37. [37]

    Measuring Massive Multitask Language Understanding

    [Online]. Available: https://arxiv.org/abs/2009.03300

    work page internal anchor Pith review Pith/arXiv arXiv 2009

  38. [38]

    Evaluating Large Language Models Trained on Code

    M. Chen, J. Tworek, H. Jun, Q. Yuan, H. P. de Oliveira Pinto, J. Kaplan, H. Edwards, Y . Burda, N. Joseph, G. Brockman, A. Ray, R. Puri, G. Krueger, M. Petrov, H. Khlaaf, G. Sastry, P. Mishkin, B. Chan, S. Gray, N. Ryder, M. Pavlov, A. Power, L. Kaiser, M. Bavarian, C. Winter, P. Tillet, F. P. Such, D. Cummings, M. Plappert, F. Chantzis, E. Barnes, A. Her...

    work page internal anchor Pith review Pith/arXiv arXiv 2021

  39. [39]

    Agentdrive: An open benchmark dataset for agentic ai reasoning with llm-generated scenarios in autonomous systems,

    M. A. Ferrag, A. Lakas, and M. Debbah, “Agentdrive: An open benchmark dataset for agentic ai reasoning with llm-generated scenarios in autonomous systems,” 2026. [Online]. Available: https: //arxiv.org/abs/2601.16964

    work page arXiv 2026

  40. [40]

    Microscopic traffic simulation using sumo,

    P. A. Lopez, M. Behrisch, L. Bieker-Walz, J. Erdmann, Y .-P. Fl ¨otter¨od, R. Hilbrich, L. L ¨ucken, J. Rummel, P. Wagner, and E. Wießner, “Microscopic traffic simulation using sumo,” inProceedings of the 21st IEEE International Conference on Intelligent Transportation Systems. IEEE, 2018, pp. 2575–2582

    work page 2018

  41. [41]

    Washington, DC: The National Academies Press, 2022

    Transportation Research Board and National Academies of Sciences, Engineering, and Medicine,Highway Capacity Manual 7th Edition: A Guide for Multimodal Mobility Analysis. Washington, DC: The National Academies Press, 2022. [Online]. Available: https://nap.nationalacademies.org/catalog/26432/ highway-capacity-manual-7th-edition-a-guide-for-multimodal-mobility

    work page 2022

  42. [42]

    Lost in the Middle: How Language Models Use Long Contexts

    N. F. Liu, K. Lin, J. Hewitt, A. Paranjape, M. Bevilacqua, F. Petroni, and P. Liang, “Lost in the middle: How language models use long contexts,”Transactions of the Association for Computational Linguistics, vol. 12, pp. 157–173, 02 2024. [Online]. Available: https://doi.org/10.1162/tacl a 00638

    work page internal anchor Pith review doi:10.1162/tacl 2024