Pith. sign in

REVIEW 4 major objections 4 minor 4 cited by

A systematic review of roughly 100 CARLA reinforcement-learning papers maps the field and finds model-free methods dominate, with over 80% of studies using DQN, PPO, SAC, DDPG, or TD3.

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 · deepseek-v4-flash

2026-08-04 20:59 UTC pith:BZMZCDYF

load-bearing objection Useful survey of CARLA RL work, but the headline numbers rest on a selection process that is never described, and the 'peer-reviewed' label does not survive contact with the reference list. the 4 major comments →

arxiv 2509.08221 v1 pith:BZMZCDYF submitted 2025-09-10 cs.RO

A Comprehensive Review of Reinforcement Learning for Autonomous Driving in the CARLA Simulator

Elahe Delavari , Feeza Khan Khanzada , Jaerock Kwon This is my paper
classification cs.RO
keywords reinforcement learningautonomous drivingCARLA simulatormodel-free RLreward designevaluation metricssim-to-real transfersurvey
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.

This paper is a survey aiming to give a complete picture of how reinforcement learning is actually used for autonomous driving inside the CARLA simulator. It claims to systematically analyze about 100 peer-reviewed papers, grouping them by algorithmic family, state representation, action space, reward design, terminal conditions, and evaluation metrics. The central quantitative finding is that model-free methods account for over 80% of the surveyed work, while model-based and hybrid approaches are rare. The authors also catalog which sensors, control abstractions, reward terms, and metrics recur, and distill persistent challenges such as sparse rewards, sim-to-real transfer, and limited safety guarantees into open research questions. A sympathetic reader would care because the paper promises a unified map that newcomers can use to orient themselves and that seasoned researchers can use to spot underexplored directions.

Core claim

The paper's central claim is that the CARLA RL literature, sampled as roughly 100 IEEE and ACM papers, can be organized into a taxonomy whose quantitative shape is currently dominated by model-free methods: DQN, PPO, SAC, DDPG, TD3, and variants together appear in more than 80% of studies. The survey further claims that state representations most often rely on front-facing RGB or kinematic vehicle states rather than bird's-eye-view or semantic images; that combined discrete or continuous action spaces are the norm; that collision penalties, speed control, and lane-deviation terms are the most frequent reward components; and that evaluation is split across at least 95 unique metrics, with suc

What carries the argument

The organizing device is a four-way algorithmic taxonomy—model-free, model-based, hierarchical, and hybrid—combined with design axes for state space (RGB, LiDAR, BEV, semantic maps, kinematics), action space (discrete, continuous, hierarchical), reward terms, terminal conditions, and evaluation metrics. The taxonomy does the work of converting a scattered body of papers into quantitative distributions, such as the over-80% model-free share, the frequency of collision-based rewards, and the concentration of experiments in a small subset of CARLA towns.

Load-bearing premise

The survey's quantitative claims rest on the assumption that the roughly 100 papers it selected from IEEE and ACM sources are representative of the peer-reviewed CARLA RL literature, but the paper does not report its search or screening protocol.

What would settle it

A reproducible literature search with explicit inclusion and exclusion criteria across the same databases would settle the central quantitative claims: if a systematically constructed corpus yields a model-free share clearly different from the reported figure, or a different count of unique evaluation metrics, the survey's map of the field would need revision.

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

Share X Bluesky LinkedIn Reddit HN

If this is right

  • If the survey's sample is representative, then model-based and hybrid RL are genuinely underused in CARLA research, suggesting a concrete opening for methods that learn environment dynamics.
  • The prevalence of kinematic and RGB state representations implies that many current policies may not exploit the spatial structure that BEV and semantic inputs provide, which could limit generalization.
  • The identification of 95 unique evaluation metrics means cross-paper comparison is currently fragmented; a move toward standardized metrics would make reported success rates and collision rates more meaningful.
  • The recurring limitations—sparse rewards, sim-to-real gap, weak safety guarantees, and narrow behavior diversity—indicate that progress in CARLA RL depends less on new model-free variants than on addressing these systemic gaps.
  • The survey's town and scenario mapping shows that most experiments concentrate on the simpler official towns, so claims of urban driving competence are often tested only in relatively easy settings.

Where Pith is reading between the lines

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

  • A reader should treat the 80% model-free figure as a provisional estimate: the paper reports no search strings, database queries, or inclusion and exclusion criteria, so the sample's representativeness cannot be independently verified from the text.
  • If the sampling premise fails, the quantitative statistics (method shares, reward frequencies, metric counts) could shift even though the individual paper summaries remain accurate; the taxonomy itself would still be a usable map.
  • The implied call for standardized CARLA benchmarks is testable: a future survey could apply a reproducible search protocol to the same databases and check whether the distributions and the 95-metric count are stable.
  • The paper leaves implicit that the same taxonomy could be applied to newer CARLA releases and leaderboard submissions, which would show whether model-free dominance is a stable property of the field or an artifact of the reviewed time window.

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

4 major / 4 minor

Summary. This manuscript is a literature review of reinforcement learning (RL) research conducted in the CARLA simulator. The authors claim to systematically analyze around 100 peer-reviewed papers and organize the field according to algorithmic family (model-free, model-based, hybrid, and alternative methods), state-space representations, action-space designs, reward functions, terminal conditions, evaluation metrics, and CARLA towns/scenarios. The paper reports quantitative prevalence, most notably that over 80% of surveyed works use model-free RL, and identifies persistent challenges and future research directions.

Significance. If the quantitative claims are reliable, the survey would provide a useful map of design choices and evaluation practices in CARLA-based RL, valuable to both newcomers and experienced researchers. The manuscript's strengths include its broad coverage, structured taxonomies in Tables 1-9, explicit discussion of limitations, and useful categorization of rewards, terminal conditions, and evaluation metrics. However, the central value of the paper is an empirical aggregation over an unreported and unverifiable sample. The stress-test concern about the undefined selection protocol is confirmed by the manuscript text: the only inclusion description is 'we consider around 100 papers from IEEE and ACM,' with no search strings, databases, inclusion/exclusion criteria, or list of included papers. As a result, the paper's quantitative conclusions are currently not testable or reproducible.

major comments (4)
  1. [Section 1 (final paragraph) and throughout] The only selection description is 'we consider around 100 papers from IEEE and ACM' (end of Section 1). No search strings, databases, date range, inclusion/exclusion criteria, screening steps, or an enumerated list of included papers are reported. Every aggregate result — the >80% model-free share (Fig. 1), the 95 unique evaluation metrics (§8), and the reward/terminal-condition/town distributions (Figs. 3-5; Tables 5, 7, 9) — depends on this unspecified sample. As written, the quantitative claims are not reproducible. Please add a Methods subsection describing the full protocol and release the list of included papers and per-paper extraction data, e.g., as a supplementary table.
  2. [Abstract / Conclusion vs. Reference list] The abstract and conclusion describe the corpus as 'around 100 peer-reviewed papers,' but numerous cited works are arXiv preprints with no verified peer-reviewed status (e.g., [14], [15], [16], [17], [45], [48], [59], [60], [72], [86], [101], [106], [113], [117], [118], [127]). Some carry comments such as 'Accepted by ECCV 2024' or 'Accepted at CVPR 2020,' but no systematic verification or recording of peer-review status is provided. The 'peer-reviewed' label is therefore unsupported. Please either restrict the corpus to verified peer-reviewed publications or explicitly report preprint status and adjust the claims accordingly.
  3. [§3.1.6 and Tables 5, 2, 3] The extraction tables contain internal errors that undermine confidence in the per-paper categorization. For example, §3.1.6 attributes [69] to 'Weng et al.' but the reference list gives 'Friji et al.'; Table 5 contains duplicated reference indices in the Speed Penalty row ('70, 70, 71, 71') and malformed entries in the Deviation and Travel rows ('101],[14]', '14],[11'); §3.1.2 refers to 'Doe et al. [36]', and reference [36] lacks a title and venue. Please audit every table entry against the cited source and correct the bibliography and in-text attributions.
  4. [§5.1 (Table 2 footnote)] The action-space taxonomy relies on an unvalidated coding rule: 'we categorized the mapping strategy of the papers that has not explicitly mentioned the low level controller as direct.' Since the direct/indirect mapping distinction is a reported finding (Tables 2 and 3), this ad-hoc rule should be stated as a coding decision, the number of papers classified under it should be reported, and a sensitivity check (e.g., second-coder agreement or a re-analysis excluding those papers) should be provided.
minor comments (4)
  1. [§4.3] Heading typo: 'Segmentic' should be 'Semantic'. Also within §4.3, 'beyesian' should be 'Bayesian'.
  2. [Figures 1-5] The figures are word-cloud/pie-style with no numeric axes, counts, or legends defining the meaning of token size/color. Since the manuscript makes quantitative claims, please add exact percentages, counts, or a supplementary table so readers can verify the displayed shares.
  3. [Table 9] The checkmark matrix of scenario types versus CARLA towns has no per-cell citations. It is impossible to tell which references support each town-scenario combination. Please add citations or provide a supplementary table mapping each included paper to the towns and scenarios used.
  4. [Reference list] Several references are incomplete or inconsistently formatted, e.g., [36], [53], [64], and [95] lack titles or venues, and [71] is formatted as an arXiv identifier. A full bibliography consistency pass is needed.

Circularity Check

0 steps flagged

No circularity: this is an aggregation survey; its quantitative claims are empirical counts over the authors' sample, not derived predictions or fitted outputs.

full rationale

This paper is a literature review and taxonomy of roughly 100 CARLA-based RL papers. It does not derive equations, fit parameters to data, or make predictions from a model. The central quantitative claim—that over 80% of surveyed works are model-free—is an empirical count over the authors' selected sample, presented as a distribution in Fig. 1. That is an aggregation of reported categories, not a quantity defined in terms of the conclusion it supports. Likewise, counts such as '95 unique evaluation metrics' and the reward/terminal-condition distributions are tabulations of the reviewed papers. No step in the paper reduces a 'prediction' to an input by construction, because there are no predictive derivations at all. I checked for self-citation load-bearing: the reference list contains no entries authored by Delavari, Khanzada, or Kwon, so the paper does not import uniqueness or ansatz from the authors' own prior work. The main weakness is sample representativeness: the authors say 'we consider around 100 papers from IEEE and ACM' without reporting search strings, inclusion/exclusion criteria, or a PRISMA-style flow, and several included items are arXiv preprints rather than verified peer-reviewed publications. That is a legitimate threat to the survey's coverage claim, but it is a selection-bias / reproducibility concern, not circularity. The absence of a reported sampling protocol does not make the survey's numerical summaries equivalent to its own inputs by definition. Therefore the appropriate circularity score is 0.

Axiom & Free-Parameter Ledger

0 free parameters · 3 axioms · 0 invented entities

This review introduces no free parameters or entities. It rests on load-bearing premises: the sampled papers are representative of the CARLA RL literature, the classification rules (including the stated assumption for unmapped controllers) are accurate, and the claimed peer-review status holds.

axioms (3)
  • domain assumption The ~100 selected papers are representative of the peer-reviewed CARLA RL literature.
    Section 1 states "we consider around 100 papers from IEEE and ACM" without a documented search or inclusion protocol; all quantitative estimates depend on this sample.
  • ad hoc to paper For papers that do not explicitly mention a low-level controller, the action mapping strategy is categorized as direct mapping.
    Footnote in Section 5.1: "In this study we categorized the mapping strategy of the papers that has not explicitly mentioned the low level controller as direct." This affects Tables 2 and 3.
  • domain assumption Papers cited as peer-reviewed are indeed peer-reviewed.
    The abstract claims peer-reviewed papers, but the bibliography includes several arXiv preprints (e.g., refs [14]-[17], [20], [48], [86]) without explicit venue verification.

pith-pipeline@v1.3.0-alltime-deepseek · 46713 in / 9991 out tokens · 104312 ms · 2026-08-04T20:59:44.325731+00:00 · methodology

0 comments
Cite this review

Pith. "Pith review of A Comprehensive Review of Reinforcement Learning for Autonomous Driving in the CARLA Simulator." pith.science (2026). https://pith.science/paper/BZMZCDYF

@misc{pith2026250908221,
  author       = {Pith},
  title        = {Pith review of: A Comprehensive Review of Reinforcement Learning for Autonomous Driving in the CARLA Simulator},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/BZMZCDYF}},
  note         = {Machine review of arXiv:2509.08221}
}
Share X Bluesky LinkedIn Reddit HN
read the original abstract

Autonomous-driving research has recently embraced deep Reinforcement Learning (RL) as a promising framework for data-driven decision making, yet a clear picture of how these algorithms are currently employed, benchmarked and evaluated is still missing. This survey fills that gap by systematically analysing around 100 peer-reviewed papers that train, test or validate RL policies inside the open-source CARLA simulator. We first categorize the literature by algorithmic family model-free, model-based, hierarchical, and hybrid and quantify their prevalence, highlighting that more than 80% of existing studies still rely on model-free methods such as DQN, PPO and SAC. Next, we explain the diverse state, action and reward formulations adopted across works, illustrating how choices of sensor modality (RGB, LiDAR, BEV, semantic maps, and carla kinematics states), control abstraction (discrete vs. continuous) and reward shaping are used across various literature. We also consolidate the evaluation landscape by listing the most common metrics (success rate, collision rate, lane deviation, driving score) and the towns, scenarios and traffic configurations used in CARLA benchmarks. Persistent challenges including sparse rewards, sim-to-real transfer, safety guarantees and limited behaviour diversity are distilled into a set of open research questions, and promising directions such as model-based RL, meta-learning and richer multi-agent simulations are outlined. By providing a unified taxonomy, quantitative statistics and a critical discussion of limitations, this review aims to serve both as a reference for newcomers and as a roadmap for advancing RL-based autonomous driving toward real-world deployment.

discussion (0)

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

Forward citations

Cited by 4 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score.

  1. Sim2Real-AD: A Modular Sim-to-Real Framework for Deploying VLM-Guided Reinforcement Learning in Real-World Autonomous Driving

    cs.RO 2026-04 unverdicted novelty 6.0

    Sim2Real-AD enables zero-shot transfer of CARLA-trained VLM-guided RL policies to full-scale vehicles, reporting 75-90% success rates in car-following, obstacle avoidance, and stop-sign scenarios without real-world RL...

  2. Zero-Human Demonstration End-to-end Autonomous Driving with Trajectory Scorer

    cs.RO 2025-10 conditional novelty 6.0

    A reward-only offline RL method for trajectory planning in end-to-end autonomous driving achieves state-of-the-art on Navhard and competitive closed-loop HUGSIM performance without imitation learning.

  3. Dreaming Across Towns: Semantic Rollout and Town-Adversarial Regularization for Zero-Shot Held-Out-Town Fixed-Route Driving in CARLA

    cs.RO 2026-04 unverdicted novelty 5.0

    Semantic rollout prediction plus town-adversarial regularization on a Dreamer agent raises mean zero-shot success rate for fixed-route driving across held-out CARLA towns under fixed weather and no traffic.

  4. Dreaming Across Towns: Semantic Rollout and Town-Adversarial Regularization for Zero-Shot Held-Out-Town Fixed-Route Driving in CARLA

    cs.RO 2026-04 unverdicted novelty 5.0

    Semantic rollout plus town-adversarial regularization raises zero-shot success in held-out CARLA towns to 36.6% and 85.6% versus matched DreamerV3 baselines.

Reference graph

Works this paper leans on

132 extracted references · 52 canonical work pages · cited by 3 Pith papers · 18 internal anchors

  1. [1]

    Tech- nical Report DOT HS 812 115, U.S

    National Highway Traffic Safety Administration: Critical reasons for crashes investigated in the national motor vehicle crash causation survey. Tech- nical Report DOT HS 812 115, U.S. Department of Transportation (2015). Accessed: 2025-04-09. https://crashstats.nhtsa.dot.gov/Api/Public/ ViewPublication/812115

    2015

  2. [2]

    Electronic Imaging29(19), 70–76 (2017) https://doi.org/10.2352/issn.2470-1173.2017.19.avm-023

    Sallab, A.E., Abdou, M., Perot, E., Yogamani, S.: Deep reinforcement learning framework for autonomous driving. Electronic Imaging29(19), 70–76 (2017) https://doi.org/10.2352/issn.2470-1173.2017.19.avm-023

    arXiv 2017

  3. [3]

    In: Proceedings of the 1st Annual Confer- ence on Robot Learning, pp

    Dosovitskiy, A., Ros, G., Codevilla, F., Lopez, A., Koltun, V.: CARLA: An Open Urban Driving Simulator. In: Proceedings of the 1st Annual Confer- ence on Robot Learning, pp. 1–16. PMLR, ??? (2017). ISSN: 2640-3498. https://proceedings.mlr.press/v78/dosovitskiy17a.htmlAccessed 2024-07-11

    2017

  4. [4]

    In: 2022 IEEE 7th International Conference on Intelligent Transportation Engineer- ing (ICITE), pp

    Wu, Y., Yin, Z., Yu, J., Zhang, M.: Lane Change Decision-Making through Deep Reinforcement Learning with Driver’s Inputs. In: 2022 IEEE 7th International Conference on Intelligent Transportation Engineer- ing (ICITE), pp. 314–319 (2022). https://doi.org/10.1109/ICITE56321.2022. 10101421 .https://ieeexplore.ieee.org/document/10101421/?arnumber=10101421 Acc...

    arXiv 2022

  5. [5]

    In: Advances in Robotics - 5th International Conference of The Robotics Society, pp

    Goel, A., Chauhan, S.: Adaptive Look-ahead distance for Pure Pursuit Controller with Deep Reinforcement Learning Techniques. In: Advances in Robotics - 5th International Conference of The Robotics Society, pp. 1–

  6. [6]

    https://doi.org/10.1145/3478586.3478600

    ACM, Kanpur India (2021). https://doi.org/10.1145/3478586.3478600 . https://dl.acm.org/doi/10.1145/3478586.3478600Accessed 2025-01-10

    arXiv 2021

  7. [7]

    In: Proceedings of the 4th Inter- national Conference on Smart City Applications, pp

    Youssef, F., Houda, B.: Deep reinforcement learning with external control: self-driving car application. In: Proceedings of the 4th Inter- national Conference on Smart City Applications, pp. 1–7. ACM, Casablanca Morocco (2019). https://doi.org/10.1145/3368756.3369038 . https://dl.acm.org/doi/10.1145/3368756.3369038Accessed 2025-01-10

    arXiv 2019

  8. [8]

    In: 2021 IEEE Winter Conference on Applica- tions of Computer Vision Workshops (W ACVW), pp

    Carton, F., Filliat, D., Rabarisoa, J., Pham, Q.C.: Using Semantic Infor- mation to Improve Generalization of Reinforcement Learning Policies for Autonomous Driving. In: 2021 IEEE Winter Conference on Applica- tions of Computer Vision Workshops (W ACVW), pp. 144–151. IEEE, Waikola, HI, USA (2021). https://doi.org/10.1109/W ACVW52041.2021.00020 . https://i...

    arXiv 2021

  9. [9]

    In: 2024 IEEE/SICE International Symposium on System Integration (SII), pp

    Mohammed, S., Argun, A., Ascheid, G.: A Unified Approach to Autonomous Driving in a High-Fidelity Simulator Using Vision- Based Reinforcement Learning. In: 2024 IEEE/SICE International Symposium on System Integration (SII), pp. 1093–1098 (2024). 52 https://doi.org/10.1109/SII58957.2024.10417385 . ISSN: 2474-2325. https://ieeexplore.ieee.org/document/10417...

    arXiv 2024

  10. [10]

    In: 2022 Tenth International Symposium on Com- puting and Networking Workshops (CANDAR W), pp

    Zhang, M., Nakamoto, Y.: Toward ensuring better learning performance in reinforcement learning. In: 2022 Tenth International Symposium on Com- puting and Networking Workshops (CANDAR W), pp. 134–139 (2022). https://doi.org/10.1109/CANDAR W57323.2022.00037 . ISSN: 2832-1324. https://ieeexplore.ieee.org/document/10062717/?arnumber=10062717Accessed 2025-01-10

    arXiv 2022

  11. [11]

    In: 2022 6th CAA International Conference on Vehicular Control and Intelligence (CVCI), pp

    Yang, Z., Pei, X., Xu, J., Zhang, X., Xi, W.: Decision-making in Autonomous Driving by Reinforcement Learning Combined with Planning & control. In: 2022 6th CAA International Conference on Vehicular Control and Intelligence (CVCI), pp. 1–6 (2022). https://doi.org/10.1109/CVCI56766.2022.9964691 .https://ieeexplore.ieee.org/document/9964691/?arnumber=996469...

    arXiv 2022

  12. [12]

    IEEE Transactions on Vehicular Tech- nology72(9), 11213–11226 (2023) https://doi.org/10.1109/TVT.2023.3266940

    Wang, J., Sun, H., Zhu, C.: Vision-Based Autonomous Driving: A Hierarchical Reinforcement Learning Approach. IEEE Transactions on Vehicular Tech- nology72(9), 11213–11226 (2023) https://doi.org/10.1109/TVT.2023.3266940 . Conference Name: IEEE Transactions on Vehicular Technology. Accessed 2025-01-10

    arXiv 2023

  13. [14]

    In: Proceedings of the Fifth International Conference on Distributed Artificial Intelligence

    Zhu, Z., Zhang, S., Zhuang, Y., Liu, Y., Liu, M., Gong, Z., Kai, S., Gu, Q., Wang, B., Cheng, S., Wang, X., Hao, J., Yu, Y.: RITA: Boost Driving Simulators with Realistic Interactive Traffic Flow. In: Proceedings of the Fifth International Conference on Distributed Artificial Intelligence. DAI ’23, pp. 1–10. Association for Computing Machinery, New York, ...

    arXiv 2023

  14. [15]

    Li, Q., Jia, X., Wang, S., Yan, J.: Think2Drive: Efficient Reinforcement Learning by Thinking in Latent World Model for Quasi-Realistic Autonomous Driving (in CARLA-v2). arXiv. Comment: Accepted by ECCV 2024 (2024). https://doi. org/10.48550/arXiv.2402.16720

    Pith/arXiv arXiv doi:10.48550/arxiv.2402.16720 2024

  15. [16]

    Pan, M., Zhu, X., Zheng, Y., Wang, Y., Yang, X.: Model-Based Reinforce- ment Learning with Isolated Imaginations. arXiv. Comment: arXiv admin note: 53 text overlap with arXiv:2205.13817 (2023). https://doi.org/10.48550/arXiv. 2303.14889

    Pith/arXiv arXiv doi:10.48550/arxiv 2023

  16. [17]

    Cai, P., Wang, H., Huang, H., Liu, Y., Liu, M.: Vision-Based Autonomous Car Racing Using Deep Imitative Reinforcement Learning. arXiv. Comment: 8 pages, 8 figures. IEEE Robotics and Automation Letters (RA-L) & IROS 2021 (2021). https://doi.org/10.48550/arXiv.2107.08325

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.48550/arxiv.2107.08325 2021

  17. [18]

    Chen, D., Koltun, V., Kr¨ ahenb¨ uhl, P.: Learning to Drive from a World on Rails. arXiv. Comment: Paper published in ICCV 2021(Oral); Code and data available at: https://dotchen.github.io/world on rails/ (2021). https://doi.org/10.48550/ arXiv.2105.00636

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.48550/arxiv.2105.00636 2021

  18. [19]

    IEEE Trans- actions on Intelligent Vehicles, 1–14 (2024) https://doi.org/10.1109/TIV.2024

    Wang, C., Zhou, S., Wang, L., Lu, Z., Wu, C., Wen, X., Shou, G.: Autonomous Driving via Knowledge-Enhanced Safe Reinforcement Learning. IEEE Trans- actions on Intelligent Vehicles, 1–14 (2024) https://doi.org/10.1109/TIV.2024. 3412916

    doi:10.1109/tiv.2024 2024

  19. [20]

    IEEE Transactions on Neural Networks and Learning Systems32(12), 5379– 5391 (2021) https://doi.org/10.1109/TNNLS.2021.3109284

    Huang, C., Zhang, R., Ouyang, M., Wei, P., Lin, J., Su, J., Lin, L.: Deduc- tive Reinforcement Learning for Visual Autonomous Urban Driving Navigation. IEEE Transactions on Neural Networks and Learning Systems32(12), 5379– 5391 (2021) https://doi.org/10.1109/TNNLS.2021.3109284

    arXiv 2021

  20. [21]

    Yurtsever, E., Capito, L., Redmill, K., Ozguner, U.: Integrating Deep Reinforce- ment Learning with Model-based Path Planners for Automated Driving. arXiv. Comment: 6 pages, 5 figures. Accepted for IEEE Intelligent Vehicles Symposium 2020 (2020). https://doi.org/10.48550/arXiv.2002.00434

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.48550/arxiv.2002.00434 2020

  21. [22]

    In: ICRA Workshop on Open Source Software, vol

    Quigley, M., Conley, K., Gerkey, B., Faust, J., Foote, T., Leibs, J., Wheeler, R., Ng, A.Y.,et al.: Ros: an open-source robot operating system. In: ICRA Workshop on Open Source Software, vol. 3, p. 5 (2009). Kobe

    2009

  22. [23]

    https://github.com/autowarefoundation/autoware

    Autoware Foundation: Autoware: Open-source software for autonomous driving. https://github.com/autowarefoundation/autoware. Accessed: June 2025

    2025

  23. [24]

    https://arxiv.org/abs/2101.05337

    Kaur, P., Taghavi, S., Tian, Z., Shi, W.: A Survey on Simulators for Testing Self-Driving Cars (2021). https://arxiv.org/abs/2101.05337

    Pith/arXiv arXiv 2021

  24. [25]

    https://arxiv.org/abs/2005.03778

    Rong, G., Shin, B.H., Tabatabaee, H., Lu, Q., Lemke, S., Moˇ zeiko, M., Boise, E., Uhm, G., Gerow, M., Mehta, S., Agafonov, E., Kim, T.H., Sterner, E., Ushiroda, K., Reyes, M., Zelenkovsky, D., Kim, S.: LGSVL Simulator: A High Fidelity Simulator for Autonomous Driving (2020). https://arxiv.org/abs/2005.03778

    Pith/arXiv arXiv 2020

  25. [26]

    In: 25th Pan-Hellenic Conference on Informatics, pp

    Chronis, C., Sardianos, C., Varlamis, I., Michail, D., Tserpes, K.: A driving profile recommender system for autonomous driving using sensor data and reinforcement learning. In: 25th Pan-Hellenic Conference on Informatics, pp. 33–38. ACM, Volos Greece (2021). https://doi.org/10.1145/3503823.3503830 . 54 https://dl.acm.org/doi/10.1145/3503823.3503830Access...

    arXiv 2021

  26. [27]

    In: 2023 IEEE International Conference on Robotics and Automation (ICRA), pp

    Udatha, S., Lyu, Y., Dolan, J.: Reinforcement learning with probabilis- tically safe control barrier functions for ramp merging. In: 2023 IEEE International Conference on Robotics and Automation (ICRA), pp. 5625–

    2023

  27. [28]

    In: Proceedings of the 31st Interna- tional Conference on Machine Learning, pp

    Silver, D., Lever, G., Heess, N., Degris, T., Wierstra, D., Riedmiller, M.: Deterministic Policy Gradient Algorithms. In: Proceedings of the 31st Interna- tional Conference on Machine Learning, pp. 387–395. PMLR, ??? (2014). ISSN: 1938-7228.https://proceedings.mlr.press/v32/silver14.htmlAccessed 2025-03-27

    2014

  28. [29]

    Lillicrap, T.P., Hunt, J.J., Pritzel, A., Heess, N., Erez, T., Tassa, Y., Silver, D., Wierstra, D.: Continuous control with deep reinforcement learning. arXiv. arXiv:1509.02971 [cs] (2019). https://doi.org/10.48550/arXiv.1509.02971 . http: //arxiv.org/abs/1509.02971 Accessed 2025-03-27

    Pith/arXiv arXiv doi:10.48550/arxiv.1509.02971 2019

  29. [30]

    In: 2020 IEEE International Conference on Robotics and Automation (ICRA), pp

    Peng, M., Gong, Z., Sun, C., Chen, L., Cao, D.: Imitative reinforcement learn- ing fusing vision and pure pursuit for self-driving. In: 2020 IEEE International Conference on Robotics and Automation (ICRA), pp. 3298–3304 (2020). IEEE

    2020

  30. [31]

    Nature 518(7540), 529–533 (2015) https://doi.org/10.1038/nature14236

    Mnih, V., Kavukcuoglu, K., Silver, D., Rusu, A.A., Veness, J., Bellemare, M.G., Graves, A., Riedmiller, M., Fidjeland, A.K., Ostrovski, G., Petersen, S., Beattie, C., Sadik, A., Antonoglou, I., King, H., Kumaran, D., Wierstra, D., Legg, S., Hassabis, D.: Human-level control through deep reinforcement learning. Nature 518(7540), 529–533 (2015) https://doi....

    doi:10.1038/nature14236 2015

  31. [32]

    In: 2022 IEEE International Conference on Real-time Computing and Robotics (RCAR), pp

    Fu, W., Li, Y., Ye, Z., Liu, Q.: Decision Making for Autonomous Driving Via Multimodal Transformer and Deep Reinforcement Learning. In: 2022 IEEE International Conference on Real-time Computing and Robotics (RCAR), pp. 481–486 (2022). https://doi.org/10.1109/RCAR54675.2022.9872180

    arXiv 2022

  32. [33]

    Multimedia Tools and Applications81(3), 3553–3576 (2022) https://doi.org/10.1007/s11042-021-11437-3

    P´ erez-Gil,´O., Barea, R., L´ opez-Guill´ en, E., Bergasa, L.M., G´ omez-Hu´ elamo, C., Guti´ errez, R., D ´ ıaz-D ´ ıaz, A.: Deep reinforcement learning based control for Autonomous Vehicles in CARLA. Multimedia Tools and Applications81(3), 3553–3576 (2022) https://doi.org/10.1007/s11042-021-11437-3

    work page doi:10.1007/s11042-021-11437-3 2022

  33. [34]

    In: ICMLCA 2021; 2nd International Conference on Machine Learning and 55 Computer Application, pp

    Zhang, B., Xu, C., Su, Y., Xu, J.: A self-learning lane keeping algorithm. In: ICMLCA 2021; 2nd International Conference on Machine Learning and 55 Computer Application, pp. 1–7 (2021)

    2021

  34. [35]

    In: 2021 IEEE International Conference on Real-time Computing and Robotics (RCAR), pp

    Chen, M., Li, Y., Liu, Q., Lv, S., Xu, Y., Liu, Y.: Towards Autonomous Driving Decision by Combining Self-attention and Deep Reinforcement Learn- ing. In: 2021 IEEE International Conference on Real-time Computing and Robotics (RCAR), pp. 1110–1115 (2021). https://doi.org/10.1109/RCAR52367. 2021.9517610

    arXiv 2021

  35. [36]

    Accessed 2025-05-06

    Doe, D.M., Chen, D., Han, K., Wang, H., Xie, J., Han, Z.: DSORL: Data source optimization with reinforcement learning scheme for vehicular named data networks24(10), 11225–11237 https://doi.org/10.1109/TITS.2023.3292033 . Accessed 2025-05-06

    arXiv 2023

  36. [37]

    In: 2023 IEEE International Symposium on Robotic and Sensors Environments (ROSE), pp

    Tsai, J., Chang, Y.-T., Chuang, P.-H., You, Z.: An Autonomous Vehicle- Following Technique for Self-Driving Cars Based on the Semantic Segmentation Technique. In: 2023 IEEE International Symposium on Robotic and Sensors Environments (ROSE), pp. 1–7 (2023). https://doi.org/10.1109/ROSE60297. 2023.10410810

    arXiv 2023

  37. [38]

    Accessed 2025-05-06

    Ahmed, M., Abobakr, A., Lim, C.P., Nahavandi, S.: Policy-based reinforcement learning for training autonomous driving agents in urban areas with affor- dance learning23(8), 12562–12571 https://doi.org/10.1109/TITS.2021.3115235 . Accessed 2025-05-06

    arXiv 2021

  38. [39]

    In: 2022 International Conference on Advanced Robotics and Mechatronics (ICARM), pp

    Li, L., Jiang, W., Shi, M., Wu, T.: Dynamic target following control for autonomous vehicles with deep reinforcement learning. In: 2022 International Conference on Advanced Robotics and Mechatronics (ICARM), pp. 386–

    2022

  39. [40]

    Schulman, J., Wolski, F., Dhariwal, P., Radford, A., Klimov, O.: Proximal Policy Optimization Algorithms. arXiv. arXiv:1707.06347 [cs] (2017). https:// doi.org/10.48550/arXiv.1707.06347 . http://arxiv.org/abs/1707.06347 Accessed 2025-03-27

    Pith/arXiv arXiv doi:10.48550/arxiv.1707.06347 2017

  40. [41]

    In: 2023 International Joint Conference on Neural Networks (IJCNN), pp

    Deng, Q., Zhao, Y., Li, R., Hu, Q., Liu, T., Li, R.: Context - Enhanced Meta-Reinforcement Learning with Data-Reused Adapta- tion for Urban Autonomous Driving. In: 2023 International Joint Conference on Neural Networks (IJCNN), pp. 1–8 (2023). https: //doi.org/10.1109/IJCNN54540.2023.10191187 . ISSN: 2161-4407. https://ieeexplore.ieee.org/document/1019118...

    arXiv 2023

  41. [42]

    In: 2021 26th International Conference on Automation and Computing (ICAC), pp

    Cheng, S., Wang, N., Chen, F., Pipe, T.: Longitudinal Driv- ing Skills Transfer from Driver to Smart Vehicle. In: 2021 26th International Conference on Automation and Computing (ICAC), pp. 1–6 (2021). https://doi.org/10.23919/ICAC50006.2021.9594177 . https://ieeexplore.ieee.org/document/9594177/?arnumber=9594177Accessed 2025-01-10

    arXiv 2021

  42. [43]

    In: 2024 IEEE 13th Data Driven Control and Learning Systems Conference (DDCLS), pp

    Zhao, J., Zhao, Y., Li, W., Zeng, C.: End-to-End Autonomous Driving Algo- rithm Based on PPO and Its Implementation. In: 2024 IEEE 13th Data Driven Control and Learning Systems Conference (DDCLS), pp. 1852–1858 (2024). https://doi.org/10.1109/DDCLS61622.2024.10606596 . ISSN: 2767-9861. https://ieeexplore.ieee.org/document/10606596/?arnumber=10606596Access...

    work page arXiv 2024

  43. [44]

    In: 2023 38th Youth Academic Annual Conference of Chinese Association of Automation (YAC), pp

    Wu, Y., Yuan, X.: Proximal Policy Optimization-based Reinforcement Learn- ing for End-to-end Autonomous Driving. In: 2023 38th Youth Academic Annual Conference of Chinese Association of Automation (YAC), pp. 844–849 (2023). https://doi.org/10.1109/YAC59482.2023.10401381 . ISSN: 2837-8601. https://ieeexplore.ieee.org/document/10401381/?arnumber=10401381Acc...

    work page arXiv 2023

  44. [45]

    IEEE Transactions on Vehicular Technology73(2), 1567–1581 (2024) https://doi.org/10.1109/TVT

    Deng, Q., Li, R., Hu, Q., Zhao, Y., Li, R.: Context-Aware Meta-RL With Two-Stage Constrained Adaptation for Urban Driving. IEEE Transactions on Vehicular Technology73(2), 1567–1581 (2024) https://doi.org/10.1109/TVT. 56 2023.3312495 . Conference Name: IEEE Transactions on Vehicular Technology. Accessed 2025-01-10

    work page arXiv 2024

  45. [46]

    In: 2021 IEEE International Intelligent Trans- portation Systems Conference (ITSC), pp

    Agarwal, T., Arora, H., Schneider, J.: Learning Urban Driving Policies using Deep Reinforcement Learning. In: 2021 IEEE International Intelligent Trans- portation Systems Conference (ITSC), pp. 607–614 (2021). https://doi.org/10. 1109/ITSC48978.2021.9564412

    arXiv 2021

  46. [47]

    Trumpp, R., B¨ uchner, M., Valada, A., Caccamo, M.: Efficient Learning of Urban Driving Policies Using Bird’s-Eye-View State Representations. arXiv. Comment: IEEE International Conference on Intelligent Transportation Systems 2023 (2023). https://doi.org/10.48550/arXiv.2305.19904

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.48550/arxiv.2305.19904 2023

  47. [48]

    Zhang, Z., Liniger, A., Dai, D., Yu, F., Gool, L.V.: End-to-End Urban Driving by Imitating a Reinforcement Learning Coach. arXiv. Comment: Published at ICCV 2021 (2021). https://doi.org/10.48550/arXiv.2108.08265

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.48550/arxiv.2108.08265 2021

  48. [49]

    In: 2021 IEEE International Conference on Image Processing (ICIP), pp

    Anzalone, L., Barra, S., Nappi, M.: Reinforced Curriculum Learning For Autonomous Driving In Carla. In: 2021 IEEE International Conference on Image Processing (ICIP), pp. 3318–3322 (2021). https://doi.org/10.1109/ICIP42928. 2021.9506673

    work page arXiv 2021

  49. [50]

    In: 2021 Latin American Robotics Symposium (LARS), 2021 Brazilian Symposium on Robotics (SBR), and 2021 Workshop on Robotics in Education (WRE), pp

    Silva, V.A.S., Grassi, V.: Addressing Lane Keeping and Intersections using Deep Conditional Reinforcement Learning. In: 2021 Latin American Robotics Symposium (LARS), 2021 Brazilian Symposium on Robotics (SBR), and 2021 Workshop on Robotics in Education (WRE), pp. 330–335 (2021). https://doi. org/10.1109/LARS/SBR/WRE54079.2021.9605436 57

    work page arXiv 2021

  50. [51]

    Xing, J., Nagata, T., Chen, K., Zou, X., Neftci, E., Krichmar, J.L.: Domain Adaptation In Reinforcement Learning Via Latent Unified State Representation. arXiv. Comment: Accepted by AAAI 2021; Fixed a typo in equation 3 (2021). https://doi.org/10.48550/arXiv.2102.05714

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.48550/arxiv.2102.05714 2021

  51. [53]

    Jin, Y.-L., Ji, Z.-Y., Zeng, D., Zhang, X.-P.: VWP:an efficient DRL-based autonomous driving model26, 2096–2108 https://doi.org/10.1109/TMM.2022. 3177942 . Accessed 2025-05-06

    work page doi:10.1109/tmm.2022 2096

  52. [54]

    In: 2023 IEEE 35th International Conference on Tools with Arti- ficial Intelligence (ICTAI), pp

    Albilani, M., Bouzeghoub, A.: Guided hierarchical reinforcement learning for safe urban driving. In: 2023 IEEE 35th International Conference on Tools with Arti- ficial Intelligence (ICTAI), pp. 746–753. https://doi.org/10.1109/ICTAI59109. 2023.00115 . ISSN: 2375-0197. https://ieeexplore.ieee.org/document/10356414 Accessed 2025-05-06

    work page arXiv 2023

  53. [55]

    In: 2024 IEEE Intelli- gent Vehicles Symposium (IV), pp

    Guti´ errez-Moreno, R., Barea, R., L´ opez-Guill´ en, E., Arango, F., Revenga, P., Bergasa, L.M.: Decision Making for Autonomous Driving Stack: Shortening the Gap from Simulation to Real-World Implementations. In: 2024 IEEE Intelli- gent Vehicles Symposium (IV), pp. 3107–3113 (2024). https://doi.org/10.1109/ IV55156.2024.10588560

    arXiv 2024

  54. [56]

    https://ieeexplore.ieee.org/document/10422486 Accessed 2025-05-06

    work page arXiv 2025

  55. [57]

    In: 2021 28th IEEE International Conference on Electronics, Circuits, and Systems (ICECS), pp

    Muhammed, A., Essam, H., Alber, B., Samuel, K., Muhammed, H., Wagdy, M., Khaled, N., Fawzy, H., Tarek, A., AbdelSalam, M., El-Kharashi, M.W.: Develop- ing AI Agent with Functional Mockup Units for Car Autonomous Navigation. In: 2021 28th IEEE International Conference on Electronics, Circuits, and Systems (ICECS), pp. 1–5 (2021). https://doi.org/10.1109/IC...

    work page arXiv 2021

  56. [59]

    arXiv (2022)

    Chekroun, R., Toromanoff, M., Hornauer, S., Moutarde, F.: GRI: General Rein- forced Imitation and Its Application to Vision-Based Autonomous Driving. arXiv (2022). https://doi.org/10.48550/arXiv.2111.08575

    Pith/arXiv arXiv doi:10.48550/arxiv.2111.08575 2022

  57. [61]

    In: 2024 2nd International Conference on Electrical Engineering and Automatic Control (ICEEAC), pp

    Marouane, C., Saad, B.: Safe Navigation Based on Deep Q-Network Algorithm Using an Improved Control Architecture. In: 2024 2nd International Conference on Electrical Engineering and Automatic Control (ICEEAC), pp. 1–6 (2024). https://doi.org/10.1109/ICEEAC61226.2024.10576248

    work page arXiv 2024

  58. [62]

    In: 2022 International Conference on Innovation and Intelligence for Informatics, Computing, and Technologies (3ICT), pp

    Elallid, B.B., Benamar, N., Mrani, N., Rachidi, T.: Dqn-based reinforcement learning for vehicle control of autonomous vehicles interacting with pedestrians. In: 2022 International Conference on Innovation and Intelligence for Informatics, Computing, and Technologies (3ICT), pp. 489–493 (2022). IEEE 58

    2022

  59. [63]

    In: 2023 International Wireless Communications and Mobile Computing (IWCMC), pp

    Elallid, B.B., Bagaa, M., Benamar, N., Mrani, N.: A Reinforcement Learning Based Approach for Controlling Autonomous Vehicles in Complex Scenarios. In: 2023 International Wireless Communications and Mobile Computing (IWCMC), pp. 1358–1364 (2023). https://doi.org/10.1109/IWCMC58020.2023.10182377

    arXiv 2023

  60. [64]

    Toromanoff, M., Wirbel, E., Moutarde, F.: End-to-End Model-Free Reinforce- ment Learning for Urban Driving Using Implicit Affordances. arXiv. Comment: Accepted at main conference of CVPR 2020 (2020). https://doi.org/10.48550/ arXiv.1911.10868

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.48550/arxiv.1911.10868 2020

  61. [65]

    In: 2021 IEEE International Conference on Systems, Man, and Cybernetics (SMC), pp

    Ahmed, M., Lim, C.P., Nahavandi, S.: A Deep Q-Network Reinforcement Learning-Based Model for Autonomous Driving. In: 2021 IEEE International Conference on Systems, Man, and Cybernetics (SMC), pp. 739–744 (2021). https://doi.org/10.1109/SMC52423.2021.9658892

    arXiv 2021

  62. [66]

    In: 2023 International Seminar on Intelligent Technology and Its Applications (ISITIA), pp

    Muhtadin, Meliaz, M.R., Dikairono, R., Purnama, I.K.E., Purnomo, M.H.: Deep Reinforcement Learning Control Strategy at Roundabout for i-CAR Autonomous Car. In: 2023 International Seminar on Intelligent Technology and Its Applications (ISITIA), pp. 473–478 (2023). https://doi.org/10.1109/ ISITIA59021.2023.10221077

    work page arXiv 2023

  63. [67]

    IEEE Sensors Journal22(5), 4687–4696 (2022) https://doi.org/10

    Li, G., Lin, S., Li, S., Qu, X.: Learning Automated Driving in Complex Inter- section Scenarios Based on Camera Sensors: A Deep Reinforcement Learning Approach. IEEE Sensors Journal22(5), 4687–4696 (2022) https://doi.org/10. 1109/JSEN.2022.3146307

    work page arXiv 2022

  64. [68]

    May, J., Poudel, S., Hamdan, S., Poudel, K., Vargas, J.: Using the CARLA Sim- ulator to Train A Deep Q Self-Driving Car to Control a Real-World Counterpart on A College Campus

  65. [69]

    In: 2020 IEEE Global Conference on Artificial Intelligence and Internet of Things (GCAIoT), pp

    Friji, H., Ghazzai, H., Besbes, H., Massoud, Y.: A DQN-Based Autonomous Car- Following Framework Using RGB-D Frames. In: 2020 IEEE Global Conference on Artificial Intelligence and Internet of Things (GCAIoT), pp. 1–6 (2020). https: //doi.org/10.1109/GCAIoT51063.2020.9345899

    work page arXiv 2020

  66. [70]

    In: 2023 24th International Symposium on Quality Elec- tronic Design (ISQED), pp

    Clemmons, J., Jin, Y.-F.: Reinforcement Learning-Based Guidance of Autonomous Vehicles. In: 2023 24th International Symposium on Quality Elec- tronic Design (ISQED), pp. 1–6 (2023). https://doi.org/10.1109/ISQED57927. 2023.10129362

    work page arXiv 2023

  67. [71]

    DeFIX: Detecting and Fixing Failure Scenarios with Reinforcement Learning in Imitation Learning Based Autonomous Driving

    [2210.16567] DeFIX: Detecting and Fixing Failure Scenarios with Rein- forcement Learning in Imitation Learning Based Autonomous Driving. https://arxiv.org/abs/2210.16567

    work page internal anchor Pith review Pith/arXiv arXiv

  68. [72]

    In: 2021 7th International Conference on Computer and Communications (ICCC), pp

    Liu, Z., Hu, J., Song, T., Huang, Z.: A Methodology Based on Deep Rein- forcement Learning to Autonomous Driving with Double Q-Learning. In: 2021 7th International Conference on Computer and Communications (ICCC), pp. 59 1266–1271 (2021). https://doi.org/10.1109/ICCC54389.2021.9674600

    work page arXiv 2021

  69. [73]

    https://dl.acm.org/doi/10.1145/3605098.3635992 Accessed 2025-05-06

    Explainable Artificial Intelligence (XAI) Approach for Reinforcement Learning Systems|Proceedings of the 39th ACM/SIGAPP Symposium on Applied Com- puting. https://dl.acm.org/doi/10.1145/3605098.3635992 Accessed 2025-05-06

    work page arXiv 2025

  70. [74]

    In: 2020 11th International Conference on Information and Communication Systems (ICICS), pp

    Hishmeh, L., Awad, F.: Deer in The Headlights: Short Term Planning via Reinforcement Learning Algorithms for Autonomous Vehicles. In: 2020 11th International Conference on Information and Communication Systems (ICICS), pp. 255–260 (2020). https://doi.org/10.1109/ICICS49469.2020.239561

    work page arXiv 2020

  71. [75]

    Neural computation3(1), 88–97 (1991)

    Pomerleau, D.A.: Efficient training of artificial neural networks for autonomous navigation. Neural computation3(1), 88–97 (1991)

    1991

  72. [76]

    Deshpande, N., Vaufreydaz, D., Spalanzani, A.: Navigation In Urban Envi- ronments Amongst Pedestrians Using Multi-Objective Deep Reinforcement Learning. arXiv. https://doi.org/10.48550/arXiv.2110.05205 . http://arxiv.org/ abs/2110.05205 Accessed 2025-05-06

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.48550/arxiv.2110.05205 2025

  73. [77]

    Advances in neural information processing systems30(2017)

    Li, Y., Song, J., Ermon, S.: Infogail: Interpretable imitation learning from visual demonstrations. Advances in neural information processing systems30(2017)

    2017

  74. [78]

    Haarnoja, T., Zhou, A., Abbeel, P., Levine, S.: Soft Actor-Critic: Off-Policy Max- imum Entropy Deep Reinforcement Learning with a Stochastic Actor. arXiv. https://doi.org/10.48550/arXiv.1801.01290 . http://arxiv.org/abs/1801.01290 Accessed 2025-07-07

    Pith/arXiv arXiv doi:10.48550/arxiv.1801.01290 2025

  75. [79]

    In: 2021 International Joint Conference on Neural Networks (IJCNN), pp

    Savari, M., Choe, Y.: Online Virtual Training in Soft Actor- Critic for Autonomous Driving. In: 2021 International Joint Conference on Neural Networks (IJCNN), pp. 1–8 (2021). https: //doi.org/10.1109/IJCNN52387.2021.9533791 . ISSN: 2161-4407. https://ieeexplore.ieee.org/document/9533791/?arnumber=9533791Accessed 60 2025-01-10

    arXiv 2021

  76. [80]

    Advances in neural information processing systems29(2016)

    Ho, J., Ermon, S.: Generative adversarial imitation learning. Advances in neural information processing systems29(2016)

    2016

  77. [81]

    In: 2023 11th RSI International Conference on Robotics and Mechatronics (ICRoM), pp

    Aghdasian, A.J., Ardakani, A.H., Aqabakee, K., Abdollahi, F.: Autonomous Driving using Residual Sensor Fusion and Deep Rein- forcement Learning. In: 2023 11th RSI International Conference on Robotics and Mechatronics (ICRoM), pp. 265–270 (2023). https: //doi.org/10.1109/ICRoM60803.2023.10412516 . ISSN: 2572-6889. https://ieeexplore.ieee.org/document/10412...

    Pith/arXiv arXiv 2023

  78. [82]

    Advances in neural information processing systems31(2018)

    Song, J., Ren, H., Sadigh, D., Ermon, S.: Multi-agent generative adversarial imitation learning. Advances in neural information processing systems31(2018)

    2018

  79. [83]

    IEEE Access9, 93152–93164 (2021) https://doi.org/10.1109/ACCESS.2021.3085142

    Kong, S.-H., Nahrendra, I.M.A., Paek, D.-H.: Enhanced Off-Policy Reinforce- ment Learning With Focused Experience Replay. IEEE Access9, 93152–93164 (2021) https://doi.org/10.1109/ACCESS.2021.3085142

    work page arXiv 2021

  80. [84]

    In: 2023 7th CAA International Conference on Vehicular Control and Intelligence (CVCI), pp

    Wu, Y., Wang, L., Lu, X., Wu, Y., Zhang, H.: Reinforcement Learning- based Autonomous Parking with Expert Demonstrations. In: 2023 7th CAA International Conference on Vehicular Control and Intelligence (CVCI), pp. 1–6 (2023). https://doi.org/10.1109/CVCI59596.2023.10397136 .https://ieeexplore.ieee.org/document/10397136/?arnumber=10397136Accessed 2025-01-10

    arXiv 2023

Showing first 80 references.