Pith. sign in

REVIEW 4 major objections 4 minor 25 references

An end-to-end reinforcement learning policy can steer an agricultural robot down crop rows using only raw 3D LiDAR, trained entirely in simulation, with 100% success on straight rows and gradual decline on curved rows.

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

2026-08-04 15:41 UTC pith:FYZGKTOG

load-bearing objection Solid sim-only proof-of-concept for end-to-end LiDAR RL row following; the sim-to-real jump is not supported. the 4 major comments →

arxiv 2509.18608 v2 pith:FYZGKTOG submitted 2025-09-23 cs.RO cs.AI

End-to-End Crop Row Navigation via LiDAR-Based Deep Reinforcement Learning

Ana Luiza Mineiro , Francisco Affonso , Marcelo Becker This is my paper
classification cs.RO cs.AI
keywords crop row navigationunder-canopy agriculturedeep reinforcement learningLiDAR perceptionvoxel downsamplingrow followingsim-to-real transferautonomous agricultural robot
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 seeks to establish that an under-canopy agricultural robot can learn to follow crop rows end-to-end: raw 3D LiDAR point clouds go directly into a reinforcement learning policy that outputs steering commands, with no labeled datasets, no hand-crafted row-detection heuristics, and no separate perception-to-control interface. The enabler is a voxel-based downsampling step that turns a point cloud into a compact 2D row map, cutting the input by 95.83%, and stacking three consecutive maps so the policy has temporal context. In simulation, the policy completes 100-meter straight rows in every trial and traverses curved rows with gradually decreasing distance as sinusoidal frequency or amplitude increases. The authors take this as evidence that row-following can be achieved without GPS, without cameras, and without manual tuning, pending real-world validation.

Core claim

The central claim is that a policy trained with proximal policy optimization, fed a history of three 30x30 voxel-flattened LiDAR row maps and rewarded for forward progress while penalized for collisions and abrupt steering, learns a row-following controller that generalizes within the training distribution. The evidence is entirely simulated: 100% success over 100 m on straight rows; 83.89 m average on the curved training configuration; and monotonic performance decline to 9.46 m at the highest tested frequency and 19.57 m at the highest tested amplitude. The authors argue this supports treating the partially observable problem as a fully observable Markov decision process with an observatio

What carries the argument

The central object is a compact row map: raw LiDAR points are voxelized at 0.1 m resolution and vertically flattened by averaging occupancy across the middle four LiDAR channels, yielding a 30x30 grid of 900 cells. A history of three consecutive row maps is the policy input, turning the partially observable problem into a fully observable approximation. The map is what makes learning feasible: the raw 7,200-point cloud overwhelms the network, while the compact map with temporal context trains successfully.

Load-bearing premise

The load-bearing premise is that the simulated crops, LiDAR returns, and robot dynamics faithfully reproduce real under-canopy fields; all results are measured inside that simulator, and real-world validation is left to future work.

What would settle it

Run the trained policy on the same physical robot in a real under-canopy field with straight and mildly curved rows. If it collides or veers out of the row within the first 10 meters while simulation rollouts succeed, the central claim fails as stated; a quantitative check would compare real distance-before-collision against the simulated 100 m straight-row and 83.89 m curved-row baselines.

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

Share X Bluesky LinkedIn Reddit HN

If this is right

  • A LiDAR-only row-following controller can be trained entirely in simulation, eliminating labeled field data and manually designed perception-to-control interfaces.
  • The 95.83% input reduction makes a 7,200-point cloud a 900-cell map that a deep reinforcement learning policy can master within the 2 million environment steps shown in training.
  • Straight-row plantations, the most common real-world layout, are navigable with 100% success over 100 m in simulation.
  • Performance degrades gradually rather than abruptly as row curvature increases, so the policy retains partial utility under moderate unseen curvature.
  • The reward design—forward progress, collision penalty, and a clipped action-smoothness penalty—is sufficient to keep the robot near the row centerline without an explicit lateral-error term.

Where Pith is reading between the lines

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

  • Because the policy reads only local row geometry, the same voxel-map architecture could extend to other structured vegetated corridors such as orchards and vineyards, though the paper does not test this.
  • The three-frame history appears to carry useful velocity information, since the single-frame policy fails on curves; a controlled study comparing history lengths would isolate how much temporal context contributes.
  • A direct sim-to-real deployment with the same voxel map and no retraining would separate simulation-fidelity issues from policy limitations; the paper explicitly leaves real-world validation to future work.
  • The monotone performance decline across frequency and amplitude could serve as a calibration curve: matching that decline on real fields would validate the simulator, and divergence would localize the sim-to-real gap.

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. The paper presents an end-to-end crop-row navigation system for under-canopy agricultural robots. A PPO policy maps a compact LiDAR-derived representation directly to angular velocity commands (with constant forward speed) in IsaacSim. Raw Velodyne VLP-16 point clouds are voxelized and flattened across height channels into 30x30 row maps, and a three-step history of these maps is used as the policy input to address partial observability. The method is evaluated entirely in simulation: straight rows are completed with 100% success, curved rows trained at frequency 1.8 Hz and amplitude 0.20 m are followed for 83.89 +- 22.72 m on average, and performance degrades as frequency or amplitude moves away from the training distribution. Ablations show that removing the observation history or the downsampling pre-processing degrades learning.

Significance. If the central simulation claims hold, the paper is a useful demonstration that a compact voxel-flattened LiDAR representation, together with a small history buffer, is sufficient for a learned PPO policy to follow crop rows with mild curvature. The held-out frequency/amplitude tests are a genuine generalization check, and the 95.83% reduction in observation size is a practical contribution. The main weaknesses are external validity and reproducibility: the paper explicitly stops at simulation, yet contains a real-world deployment sentence, and a central reward parameter is unreported. The method is not compared to any existing row-following baseline. With corrections and tempered claims, the contribution would be a solid systems-oriented result.

major comments (4)
  1. [Section IV-B, IV-F] The paper motivates the work by real-world GNSS and lighting failures, and Section IV-B asserts that the TerraSentia platform and plantation scenes were 'accurately modeled' and 'realistic.' No comparison to real LiDAR data, real crop geometry, or real platform dynamics is provided. Section IV-F then states that 'the practicality of the trained policy for real-world deployment' is highlighted by the straight-row result. Since all evidence is from IsaacSim and Section V lists real-world validation as future work, this deployment claim is unsupported. Please either remove/temper the deployment sentence or include a concrete sim-to-real check (e.g., a real-field point cloud or a qualitative platform trial).
  2. [Section III-C, Table I] The action-smoothing penalty is defined as clip(1 - ||omega_k - omega_{k-1}||_2 / sigma, [0,1]), but the scaling constant sigma is never reported. This is a free parameter of the reward that directly shapes the learned behaviors; without its value the training setup is not fully reproducible, and the sensitivity of the results to sigma is unknown. Please report sigma and, ideally, a small sensitivity study.
  3. [Section IV-D, IV-F] It is unclear whether the straight-row configuration is part of the training distribution or a held-out geometry. Section IV-D says policies were trained 'in a sinusoidal plantation environment,' yet Section IV-F calls the straight plantation a 'baseline configuration' and Figure 5 says both scenarios 'correspond to the training environment settings.' If straight rows were not seen in training, the 100% straight-row success is a generalization result and should be explicitly described as such; if they were seen, the claim should be qualified accordingly. This affects the interpretation of the abstract's headline result.
  4. [Section III-D, Eq. (5)] Eq. (5) is inconsistent with its verbal description. The text says the vertical axis is flattened by 'averaging the occupancy across all H height levels,' but the formula is o_{x,y} = (1/H) sum_{v_i} z_{v_i}, where z_{v_i} is the vertical voxel coordinate of a voxelized point. This computes an average z-coordinate (a height/depth map), not an average occupancy, and it double-counts voxels that contain multiple points. Please correct either the formula or the description; this is a central component of the observation pipeline and must be unambiguous.
minor comments (4)
  1. [Sections I, III-D] The paper repeatedly says the policy maps 'raw 3D LiDAR data' to control commands, but the input goes through a hand-specified voxelization, channel selection, and flattening step before the policy. Consider rephrasing to 'directly from 3D LiDAR with a fixed pre-processing transform' to avoid overstating the end-to-end claim.
  2. [Figure 6] The training-return curves show single trajectories with no seeds or confidence intervals. The conclusion that 'without downsampling ... leads to failed learning' rests on one run. Please report multiple seeds or error bars, or state the number of seeds.
  3. [Table III] No comparison to existing row-following methods is provided. A geometric row detector (e.g., [8]) or a self-supervised baseline (e.g., [16]) would help the reader judge whether the learned end-to-end policy offers an actual advantage over the current state of the art, rather than just feasibility.
  4. [Section IV-C] The 95.83% reduction figure is correct only if one counts each 3D point as three scalar values (7,200*3 = 21,600 values vs. 900 cells). If the intent is to compare point count to cell count, the reduction is 87.5%. Please state the basis for the claim.

Circularity Check

0 steps flagged

No significant circularity: simulation training/evaluation is self-contained; self-citations are background only.

full rationale

The paper's central claim—that a PPO policy maps 30×30 flattened voxel row maps derived from 3D LiDAR to angular velocity—is supported by its own training/evaluation pipeline and does not reduce to its inputs. The LiDAR downsampling is a fixed preprocessing prescription (Eq. 4–5), not a fitted predictor; the 7200-to-900 size reduction is a direct arithmetic consequence of the chosen 30×30 grid, not a validated prediction. Reward components (Table I) are stated with fixed weights and are not retrofitted to held-out outcomes. The generalization experiments (Table III) use frequencies (2.0–2.6 Hz) and amplitudes (0.21–0.24 m) not present in the training environment (1.8 Hz, 0.20 m), so the observed monotone performance decline is an out-of-training-distribution result, not a forced fit. Self-citations (Refs. [3], [7], [8], [15], [16]) appear only in the related-work review and are not load-bearing for the MDP formulation, reward design, or PPO training. The paper's acknowledged sim-only limitation (Section V: 'we aim to validate our system ... in real-world environments') is a threat to external validity, not circular reasoning. No fitted parameter is relabeled as a prediction, and no equation is defined in terms of the result it purportedly establishes.

Axiom & Free-Parameter Ledger

7 free parameters · 5 axioms · 0 invented entities

The central result depends on a small set of hand-chosen observation and reward parameters plus the fidelity of the IsaacSim environment. None of these are derived or swept; they are engineering choices. The most consequential choices are the training row curvature and the LiDAR channel selection, because they define the distribution the policy learns.

free parameters (7)
  • Reward weights (r_task, r_penalty, r_collision) = 5.0, 1.0, 1.0
    Chosen manually in Table I; no sensitivity analysis is reported.
  • Action-smoothing scale sigma = not reported
    Appears in Table I's r_penalty formula but its value is never given in the paper.
  • Voxel resolution delta = 0.1 m (cubic)
    Set in Section IV-C; affects which structure survives downsampling.
  • Region of interest size = 30 x 30 voxels (3 m x 3 m)
    Fixed field around robot in Section IV-C.
  • Observation history length tau = 3
    Chosen in Section IV-D; ablation shows history matters, but no sweep over tau.
  • LiDAR channel selection H = 4 middle channels
    Selected because 'majority of hits on the plantation are concentrated' (Section IV-C); a design choice.
  • Training row curvature = 1.8 Hz, 0.20 m
    Training distribution chosen in Section IV-D; generalization is measured relative to it.
axioms (5)
  • domain assumption IsaacSim with custom assets faithfully represents TerraSentia and under-canopy crops
    Section IV-B; all results inherit this assumption and no real-world comparison is provided.
  • domain assumption A finite three-step history makes the POMDP approximately an MDP
    Section III-A; no formal bound is given, and tau=3 is chosen empirically.
  • ad hoc to paper Middle four LiDAR channels contain the row structure needed for control
    Section IV-C; points outside are discarded, and the selection is asserted rather than derived.
  • ad hoc to paper Training on one sinusoidal row geometry transfers to other amplitudes and frequencies
    Section IV-F; this extrapolation is the claimed generalization and is not guaranteed by the method.
  • domain assumption Constant forward velocity and steering-only control fully capture row-following
    Section III-B; the non-holonomic model may not hold for sharp curves, which the paper acknowledges in its performance decline.

reviewed 2026-08-04 · how reviews work

0 comments
Cite this review

Pith. "Pith review of End-to-End Crop Row Navigation via LiDAR-Based Deep Reinforcement Learning." pith.science (2026). https://pith.science/paper/FYZGKTOG

@misc{pith2026250918608,
  author       = {Pith},
  title        = {Pith review of: End-to-End Crop Row Navigation via LiDAR-Based Deep Reinforcement Learning},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/FYZGKTOG}},
  note         = {Machine review of arXiv:2509.18608}
}
Share X Bluesky LinkedIn Reddit HN
read the original abstract

Reliable navigation in under-canopy agricultural environments remains a challenge due to GNSS unreliability, cluttered rows, and variable lighting. To address these limitations, we present an end-to-end learning-based navigation system that maps raw 3D LiDAR data directly to control commands using a deep reinforcement learning policy trained entirely in simulation. Our method includes a voxel-based downsampling strategy that reduces LiDAR input size by 95.83%, enabling efficient policy learning without relying on labeled datasets or manually designed control interfaces. The policy was validated in simulation, achieving a 100% success rate in straight-row plantations and showing a gradual decline in performance as row curvature increased, tested across varying sinusoidal frequencies and amplitudes.

Figures

Figures reproduced from arXiv: 2509.18608 by Ana Luiza Mineiro, Francisco Affonso, Marcelo Becker.

Figure 1
Figure 1. Figure 1: TerraSentia robot navigating in agricultural environments [13]. [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: Workflow of the end-to-end row-following navigation system. First, raw LiDAR point clouds captured from the environment are downsampled and [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: Illustration of the proposed data downsampling method. [PITH_FULL_IMAGE:figures/full_fig_p003_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: Representation of the downsampling method. Top: LiDAR point clouds captured over time. Bottom: corresponding row maps used as policy input. [PITH_FULL_IMAGE:figures/full_fig_p004_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: Navigation trajectories of the trained policies bounded by plantation geometry. Top: Straight-row scenarios. Bottom: Curved-row scenarios. [PITH_FULL_IMAGE:figures/full_fig_p005_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: illustrates the training returns over environment steps for the different policy configurations. These compar￾isons highlight the influence of temporal context and input size on learning performance and policy robustness. Without downsampling, the high-dimensional input overwhelms the network, leading to failed learning. A single row map allows learning to begin, but including a history of observations sig… view at source ↗
Figure 7
Figure 7. Figure 7: Distance navigated at different sinusoidal row frequencies with [PITH_FULL_IMAGE:figures/full_fig_p006_7.png] view at source ↗

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

25 extracted references · 4 linked inside Pith

  1. [1]

    Tracking error learning control for precise mobile robot path tracking in outdoor environment,

    E. Kayacan and G. Chowdhary, “Tracking error learning control for precise mobile robot path tracking in outdoor environment,”Journal of Intelligent & Robotic Systems, vol. 95, pp. 975–986, 2019

    2019

  2. [2]

    Breaking the field phenotyping bottleneck in maize with autonomous robots,

    J. DeBruin, T. Aref, S. Tirado Tolosa, R. Hensley, H. Underwood, M. McGuire, C. Soman, G. Nystrom, E. Parkinson, C. Liet al., “Breaking the field phenotyping bottleneck in maize with autonomous robots,”Communications biology, vol. 8, no. 1, p. 467, 2025

    2025

  3. [3]

    Cropnav: a framework for au- tonomous navigation in real farms,

    M. V . Gasparino, V . A. Higuti, A. N. Sivakumar, A. E. Velasquez, M. Becker, and G. Chowdhary, “Cropnav: a framework for au- tonomous navigation in real farms,” in2023 IEEE International Conference on Robotics and Automation (ICRA). IEEE, 2023, pp. 11 824–11 830

    2023

  4. [4]

    Demonstrating cropfollow++: Robust under-canopy navigation with keypoints,

    A. N. Sivakumar, M. V . Gasparino, M. McGuire, V . A. H. Higuti, M. U. Akcal, and G. Chowdhary, “Demonstrating cropfollow++: Robust under-canopy navigation with keypoints,”Proceedings of Robotics: Science and Systems, Delft, Netherlands, 2024

    2024

  5. [5]

    End-to-end learning for autonomous crop row-following,

    M. Bakken, R. J. Moore, and P. From, “End-to-end learning for autonomous crop row-following,”IF AC-PapersOnLine, vol. 52, no. 30, pp. 102–107, 2019

    2019

  6. [6]

    Using a depth camera for crop row detection and mapping for under-canopy navigation of agricultural robotic vehicle,

    J. Gai, L. Xiang, and L. Tang, “Using a depth camera for crop row detection and mapping for under-canopy navigation of agricultural robotic vehicle,”Computers and Electronics in Agriculture, vol. 188, p. 106301, 2021

    2021

  7. [7]

    Multi-sensor fusion based robust row following for compact agricultural robots,

    A. E. B. Velasquez, V . A. H. Higuti, M. V . Gasparino, A. N. V . Sivakumar, M. Becker, and G. Chowdhary, “Multi-sensor fusion based robust row following for compact agricultural robots,”Field Robotics, vol. 2, pp. 1291–1319, 2022

    2022

  8. [8]

    Under canopy light detection and ranging-based autonomous navigation,

    V . A. Higuti, A. E. Velasquez, D. V . Magalhaes, M. Becker, and G. Chowdhary, “Under canopy light detection and ranging-based autonomous navigation,”Journal of Field Robotics, vol. 36, no. 3, pp. 547–567, 2019

    2019

  9. [9]

    Quantifying the lidar sim-to-real domain shift: A detailed investigation using object detectors and analyzing point clouds at target-level,

    S. Huch, L. Scalerandi, E. Rivera, and M. Lienkamp, “Quantifying the lidar sim-to-real domain shift: A detailed investigation using object detectors and analyzing point clouds at target-level,”IEEE Transactions on Intelligent V ehicles, vol. 8, no. 4, pp. 2970–2982, 2023

    2023

  10. [10]

    Localization for precision navigation in agricultural fields—beyond crop row following,

    W. Winterhalter, F. Fleckenstein, C. Dornhege, and W. Burgard, “Localization for precision navigation in agricultural fields—beyond crop row following,”Journal of Field Robotics, vol. 38, no. 3, pp. 429–451, 2021

    2021

  11. [11]

    Autonomous navigation of 4wis4wid agricultural field mobile robot using deep reinforcement learning,

    T. Baby, M. K. Gohil, and B. Bhattacharya, “Autonomous navigation of 4wis4wid agricultural field mobile robot using deep reinforcement learning,”arXiv preprint arXiv:2412.18865, 2024

    Pith/arXiv arXiv 2024

  12. [12]

    An end-to-end learning- based row-following system for an agricultural robot in structured apple orchards,

    P. Huang, L. Zhu, Z. Zhang, and C. Yang, “An end-to-end learning- based row-following system for an agricultural robot in structured apple orchards,”Mathematical Problems in Engineering, vol. 2021, no. 1, p. 6221119, 2021

    2021

  13. [13]

    of Illinois College of ACES

    U. of Illinois College of ACES. (2020) Terrasentia robots by agricultural and biological engineering faculty featured in new york times. Accessed: 2025-07-20. [Online]. Available: https://aces.illinois.edu/news/terrasentia-robots-agricultural- and-biological-engineering-faculty-featured-new-york-times

    2020

  14. [14]

    Learned visual navigation for under- canopy agricultural robots,

    A. Sivakumar, S. Modi, M. Gasparino, C. Ellis, A. Velasquez, G. Chowdhary, and S. Gupta, “Learned visual navigation for under- canopy agricultural robots,” inRobotics: Science and Systems (RSS), 2021

    2021

  15. [15]

    Navigating with finesse: Leveraging neural network-based lidar per- ception and ilqr control for intelligent agriculture robotics,

    F. A. Pinto, F. A. G. Tommaselli, M. V . Gasparino, and M. Becker, “Navigating with finesse: Leveraging neural network-based lidar per- ception and ilqr control for intelligent agriculture robotics,” in2023 Latin American Robotics Symposium (LARS), 2023 Brazilian Sympo- sium on Robotics (SBR), and 2023 Workshop on Robotics in Education (WRE). IEEE, 2023, ...

    2023

  16. [16]

    Crow: A self-supervised crop row navigation algorithm for agricultural fields,

    F. Affonso, F. A. G. Tommaselli, G. Capezzuto, M. V . Gasparino, G. Chowdhary, and M. Becker, “Crow: A self-supervised crop row navigation algorithm for agricultural fields,”Journal of Intelligent & Robotic Systems, vol. 111, no. 1, p. 28, 2025

    2025

  17. [17]

    Lifelike agility and play in quadrupedal robots using reinforcement learning and generative pre- trained models,

    L. Han, Q. Zhu, J. Sheng, C. Zhang, T. Li, Y . Zhang, H. Zhang, Y . Liu, C. Zhou, R. Zhaoet al., “Lifelike agility and play in quadrupedal robots using reinforcement learning and generative pre- trained models,”Nature Machine Intelligence, vol. 6, no. 7, pp. 787– 798, 2024

    2024

  18. [18]

    Goal-driven autonomous exploration through deep reinforcement learning,

    R. Cimurs, I. H. Suh, and J. H. Lee, “Goal-driven autonomous exploration through deep reinforcement learning,”IEEE Robotics and Automation Letters, vol. 7, no. 2, pp. 730–737, 2021

    2021

  19. [19]

    Learning robust perceptive locomotion for quadrupedal robots in the wild,

    T. Miki, J. Lee, J. Hwangbo, L. Wellhausen, V . Koltun, and M. Hutter, “Learning robust perceptive locomotion for quadrupedal robots in the wild,”Science robotics, vol. 7, no. 62, p. eabk2822, 2022

    2022

  20. [20]

    Learning quadrupedal locomotion over challenging terrain,

    J. Lee, J. Hwangbo, L. Wellhausen, V . Koltun, and M. Hutter, “Learning quadrupedal locomotion over challenging terrain,”Science robotics, vol. 5, no. 47, p. eabc5986, 2020

    2020

  21. [21]

    Carl: Learning scalable planning policies with simple rewards,

    B. Jaeger, D. Dauner, J. Beißwenger, S. Gerstenecker, K. Chitta, and A. Geiger, “Carl: Learning scalable planning policies with simple rewards,”arXiv preprint arXiv:2504.17838, 2025

    Pith/arXiv arXiv 2025

  22. [22]

    Proximal policy optimization algorithms,

    J. Schulman, F. Wolski, P. Dhariwal, A. Radford, and O. Klimov, “Proximal policy optimization algorithms,”arXiv preprint arXiv:1707.06347, 2017

    Pith/arXiv arXiv 2017

  23. [23]

    Embedded high preci- sion control and corn stand counting algorithms for an ultra-compact 3d printed field robot

    E. Kayacan, Z.-Z. Zhang, and G. Chowdhary, “Embedded high preci- sion control and corn stand counting algorithms for an ultra-compact 3d printed field robot.” inRobotics: science and systems, vol. 14, 2018, p. 9

    2018

  24. [24]

    Orbit: A unified simulation framework for interactive robot learning environments,

    M. Mittal, C. Yu, Q. Yu, J. Liu, N. Rudin, D. Hoeller, J. L. Yuan, R. Singh, Y . Guo, H. Mazhar, A. Mandlekar, B. Babich, G. State, M. Hutter, and A. Garg, “Orbit: A unified simulation framework for interactive robot learning environments,”IEEE Robotics and Automa- tion Letters, vol. 8, no. 6, pp. 3740–3747, 2023

    2023

  25. [25]

    Rsl-rl: A learning library for robotics research,

    C. Schwarke, M. Mittal, N. Rudin, D. Hoeller, and M. Hutter, “Rsl-rl: A learning library for robotics research,”arXiv preprint arXiv:2509.10771, 2025

    Pith/arXiv arXiv 2025

This paper was first reviewed by deepseek-v4-flash on August 4, 2026.