Pith. sign in

REVIEW 4 major objections 5 minor 34 references

RVN-Bench makes the case that reactive visual navigation—reaching sequential goals in unseen indoor scenes from camera images alone, with no map—should be judged by collisions as much as by goal success, and its experiments show the task is

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

2026-08-02 18:57 UTC pith:NDNBS2IH

load-bearing objection Useful collision-aware indoor navigation benchmark, but the 'visual-only' claim is undercut by a ground-truth goal vector, and one key table number doesn't reproduce. the 4 major comments →

arxiv 2603.03953 v2 pith:NDNBS2IH submitted 2026-03-04 cs.RO cs.AIcs.CV

RVN-Bench: A Benchmark for Reactive Visual Navigation

Jaewon Lee , Jaeseok Heo , Gunmin Lee , Howoong Jun , Jeongwoo Oh , Songhwai Oh This is my paper
classification cs.RO cs.AIcs.CV
keywords reactive visual navigationcollision-aware benchmarkindoor mobile robotssequential goal navigationreinforcement learningimitation learningnegative trajectory datasetsim-to-real transfer
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.

RVN-Bench tries to establish that reactive visual navigation—reaching a sequence of goals in a never-seen indoor space using only camera images and no map—can be measured as a collision-aware problem, and that this formulation is learnable. The paper argues that existing indoor visual-navigation benchmarks ignore collisions and outdoor benchmarks skip indoor clutter, so a standardized simulator with collision detection is needed. Its experiments show that reinforcement-learned policies trained in the benchmark generalize to unseen scenes, with the best baseline succeeding on about 93% of first goals and incurring 3.6 collisions per kilometer on test scenes, and that a model trained on simulated data transfers to a real indoor robot better than one trained on real data alone. This gives the community a common testbed for safety-aware navigation and a way to generate collision-ending 'negative' trajectory data that is costly to collect in the real world.

Core claim

The paper's central claim is that safe indoor navigation should be evaluated as a reactive, collision-aware sequential-goal task: at each episode the agent gets a stack of RGB frames and a relative goal vector (x_g, y_g), picks from move-forward/turn-left/turn-right/stop, and must reach multiple goals in an unseen scene, with the episode terminating on collision, timeout, or N goals. To support this, RVN-Bench provides an RL training environment, a trajectory-image dataset generator, and, distinctively, a negative-trajectory generator that manufactures collision-ending trajectories by padding the occupancy map less than the robot radius. The experiments show the best model—a distributed RL a

What carries the argument

The load-bearing object is the benchmark's collision-aware sequential-goal environment: a high-fidelity indoor simulator with discretized nonholonomic actions, a cylindrical robot model, and collision detection via a precomputed navigation mesh (a geometric walkability constraint) that restricts the endpoint of each forward step. The paired mechanisms are the relative-goal observation, given to the agent at every step, and the two dataset generators: expert trajectories come from shortest paths on an occupancy map padded beyond the robot radius, while negative trajectories come from paths padded below the radius so the agent collides; recording pre- and post-collision frames yields data that

Load-bearing premise

In the task definition, the load-bearing premise is that the agent is handed the goal position relative to its current pose at every step—perfect localization of both robot and target—even though the benchmark is described as visual-only navigation.

What would settle it

A concrete check would be to re-run the reported evaluations while replacing the exact relative goal vector with one corrupted by realistic odometry noise or estimated by a vision-based goal detector; if first-goal success and CPK degrade sharply on the same held-out scenes, the benchmark's scores are being carried by the goal oracle rather than by visual navigation.

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

If this is right

  • If RVN-Bench is taken as a standard testbed, navigation research can compare methods on safety—collisions per kilometer—rather than only on reaching goals, closing a gap the paper identifies in existing indoor benchmarks.
  • Simulation-collected data appears to substitute for expensive real-world collision data: a model trained only on RVN-Bench-generated data outperformed a model trained on 54 hours of real navigation data in real-world first-goal success, and combining datasets helped further.
  • Negative trajectories, which end in collisions and are hard or unsafe to collect on real robots, can be generated cheaply in simulation and improve imitation-learning agents when used as an auxiliary dataset.
  • Predicted monocular depth markedly improves collision avoidance, and ground-truth depth improves it further, so continued progress in depth estimation should directly translate into safer reactive navigation.
  • The benchmark leaves the collision-aware task open: even the strongest baseline reaches only about 93% first-goal success and 3.6 CPK on test scenes, so there is measurable headroom for future policies.

Where Pith is reading between the lines

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

  • The paper leaves implicit that the relative goal vector is a privileged signal; if a real robot has to infer the goal from vision or odometry, first-goal success and collision rates would likely change, so the benchmark's 'reactive visual navigation' claim should be read as 'vision-based obstacle avoidance with a goal oracle.'
  • Because collision detection is a geometric navigation-mesh constraint rather than physics-based contact, collisions with lightweight, pushable, or deformable objects are not represented; the benchmark's numbers therefore describe static-scene geometry, and real-world clutter may differ.
  • The benchmark's configurable robot footprint, camera parameters, and action discretization make it straightforward to extend beyond the single platform and static scenes that the paper names as future work; the same structure could support dynamic obstacles or a goal-image variant that removes the oracle goal vector.
  • The negative-trajectory mechanism is a transferable idea: deliberately generate failing trajectories, train an expert policy and a 'bad' policy, then use the contrast to reject unsafe actions at inference; the paper applies it only to navigation, but the scheme could be used in other imitation-learning settings where failures are expensive.

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 / 5 minor

Summary. The paper introduces RVN-Bench, a collision-aware benchmark for indoor reactive visual navigation built on Habitat 2.0 and the HM3D dataset. The benchmark defines a sequential PointGoal-style task in which an agent receives RGB observations and a relative goal vector, must reach up to 32 goal positions while avoiding collisions, and is evaluated with success rate, expected number of goals reached, and collisions per kilometer. The paper also provides an RL environment, a trajectory-image dataset generator including negative collision trajectories, and a new baseline NoMaD-Neg. Experiments compare IL, Safe-RL, and RL baselines, reporting that DDPPO-DA V2 performs best, and that a NoMaD-PointGoal model trained on simulated data transfers to a Jackal UGV in 20 real-world episodes.

Significance. If the claims hold, RVN-Bench is a useful contribution: it is a large-scale (800 train / 50 val / 50 test scenes) collision-aware indoor visual navigation benchmark with standardized metrics, RL training, and offline dataset generation, plus a public code release. The paper also provides a concrete negative-data mechanism and an initial sim-to-real study, which are valuable for the community. The benchmark is empirical and I did not find circular derivations of reported metrics from fitted parameters. However, the central 'only visual observations' framing is not supported by the task definition, and two reporting issues weaken the experimental support. These are fixable but must be addressed before the paper can be accepted.

major comments (4)
  1. [Abstract and Section III-A] The abstract and Section III-A state that the agent navigates 'using only visual observations', but the task gives the agent a privileged goal vector P_t = (x_g, y_g) relative to its current pose. This is typically provided by GPS+compass in Habitat PointGoal, not by vision. The manuscript never specifies how P_t is obtained in simulation or in the real-world Jackal experiments (Section IV-E), where NoMaD-PointGoal also requires this input. As written, RVN-Bench evaluates reactive obstacle avoidance under a perfect goal-localization oracle, not end-to-end visual navigation. Please either remove the 'only visual observations' claim, explicitly model P_t as an additional input modality, or specify the sensor/mechanism and discuss the implications for sim-to-real transfer.
  2. [Tables II and III] The same configuration is reported inconsistently. Table II lists 'DDPPO-DA V2' with Train E(G)=20.79, while Table III lists 'RGB + Predicted Depth' for DD-PPO with Train E(G)=22.70; other columns (validation/test E(G), CPK, SR1) are identical. One of these is a typo, but the discrepancy undermines the reproducibility of the main quantitative comparison. Correct the inconsistency and state which value is used in the main analysis.
  3. [Section IV-E and Table IV] The sim-to-real conclusion is based on 20 episodes (15 office, 5 house) with no error bars, confidence intervals, or per-environment breakdown. With n=20, the differences between Sim-only (SR1=0.60) and Real+Sim (SR1=0.75) are not shown to be statistically meaningful, and the CPK values (196–223) likely have large variance. Please report the raw per-episode results or confidence intervals, and soften the '3.5x/4.3x improvement' claims accordingly.
  4. [Section III-A and III-B] The task definition says the goal is 'randomly selected under certain rules' and the geodesic distance is in [d_min, d_max], but it does not specify how the goal position is sampled relative to obstacles or whether the same goal-sampling procedure is used in training and evaluation. Since the benchmark's purpose is standardized comparison, the exact sampling protocol (e.g., rejection sampling, minimum clearance, or uniform over navigable positions) should be stated precisely.
minor comments (5)
  1. [Section IV-E] The real-world text uses N_g = 5 while the benchmark uses N_goal = 32; use consistent notation.
  2. [Section III-C] The statement that NoMaD-Neg 'replaces 6.25% of expert trajectories with negative ones' is vague; specify the absolute counts for reproducibility.
  3. [Equation (1)] The CoR formula is cited from [27] but the displayed equation is missing an explicit division operator in the denominator; also define the value of alpha used in the experiments.
  4. [Figure 2] The figure caption and labels are difficult to parse; ensure all symbols (t_i, t_f, k_pre, k_post) are defined in the caption or the surrounding text.
  5. [Section IV-B] The RL baselines are said to use 'the same hyperparameters as reported in [16]', but for a benchmark it would be helpful to list the key hyperparameters (learning rate, batch size, number of timesteps, seeds) in the appendix or a table.

Circularity Check

0 steps flagged

Empirical benchmark paper; no predicted quantity reduces to a fitted input or to a load-bearing self-citation.

full rationale

RVN-Bench is a benchmark contribution, not a derivation. Its headline claims - that it provides a collision-aware visual navigation task and that policies trained on it generalize - are supported by held-out test-scene results (Table II) and a real-world Jackal evaluation (Table IV), i.e., measurements on data not used for fitting. There is no equation in which a predicted output equals an input by construction. The single self-citation is the CoR selection formula from the authors' MixGAIL paper [27], used in the secondary NoMaD-Neg baseline (Eq. 1-2); it is a published external formula with its own stated form, is not fitted to RVN-Bench data, and NoMaD-Neg is not the load-bearing evidence for the benchmark's validity, so this does not rise to circularity. The unspecified alpha in Eq. (1) and the privileged relative-goal vector P_t in Sec. III-A are reproducibility/scope issues for the 'visual-only' claim, not circular reductions: the task definition states what input the agent receives, and evaluation measures behavior under that input. The paper's own limitation statement in Sec. V (static environments, single platform) is a scoping note, not a concealed circular step. No quoted reduction can be exhibited, so the honest finding is no significant circularity.

Axiom & Free-Parameter Ledger

5 free parameters · 4 axioms · 0 invented entities

The benchmark introduces no new physical entities. Its central claim rests on modeling choices (NavMesh collision, ground-truth goal localization, simulator fidelity) and a handful of hand-chosen constants (rewards, window sizes, sampling count, unspecified alpha and negative-margin). These are standard for simulation benchmarks but should be stated explicitly.

free parameters (5)
  • Reward weights r_goal, r_collision, step penalty = 1.0, -0.1, -0.01
    Hand-chosen in Section III-B; the RL results depend on these constants.
  • Negative-trajectory window k_pre, k_post = 8, 6
    Section III-C; sets how many frames before/after a collision are recorded; affects NoMaD-Neg training.
  • NoMaD-Neg sampling count k = 8
    Section IV-B; number of trajectories sampled from each diffusion model for CoR selection.
  • CoR exponent alpha (α) = unspecified
    Eq. (1) uses α from MixGAIL but the value is not given; selection behavior depends on it.
  • Negative-data padding margin = unspecified ('smaller than agent radius')
    Section III-C; controls how deeply the unsafe path enters obstacles and changes the collision distribution.
axioms (4)
  • domain assumption NavMesh-based collision detection with hard kinematic constraint accurately captures real collisions with static obstacles.
    Section III-A: 'no physics-based contact dynamics are simulated.' Thin or unmodeled obstacles may not be detected.
  • domain assumption The agent has access to the goal position relative to its current pose at every step (ground-truth localization).
    Section III-A: 'the agent is given the goal position specified relative to its current pose.' No sensor model is provided; the task is not purely visual.
  • domain assumption HM3D scenes and Habitat 2.0 rendering are sufficiently realistic that policies trained in RVN-Bench transfer to real indoor environments.
    Section IV-E claims promising sim-to-real transfer from 20 real episodes; this assumes simulator fidelity.
  • standard math A* on ground-truth occupancy with padded margin produces expert trajectories that are safe to follow with discrete actions.
    Section III-C; standard path-planning assumption, but safety depends on padding choice.

pith-pipeline@v1.3.0-alltime-deepseek · 12025 in / 15426 out tokens · 143165 ms · 2026-08-02T18:57:39.046028+00:00 · methodology

0 comments
read the original abstract

Safe visual navigation is critical for indoor mobile robots operating in cluttered environments. Existing benchmarks, however, often neglect collisions or are designed for outdoor scenarios, making them unsuitable for indoor visual navigation. To address this limitation, we introduce the reactive visual navigation benchmark (RVN-Bench), a collision-aware benchmark for indoor mobile robots. In RVN-Bench, an agent must reach sequential goal positions in previously unseen environments using only visual observations and no prior map, while avoiding collisions. Built on the Habitat 2.0 simulator and leveraging high-fidelity HM3D scenes, RVN-Bench provides large-scale, diverse indoor environments, defines a collision-aware navigation task and evaluation metrics, and offers tools for standardized training and benchmarking. RVN-Bench supports both online and offline learning by offering an environment for online reinforcement learning, a trajectory image dataset generator, and tools for producing negative trajectory image datasets that capture collision events. Evaluations demonstrate that policies trained on RVN-Bench generalize effectively across unseen simulated environments. Furthermore, initial physical experiments using a Jackal UGV indicate promising sim-to-real transfer. Code and additional materials are available at: https://sequor-robotics-research.github.io/projects/RVN-Bench/.

Figures

Figures reproduced from arXiv: 2603.03953 by Gunmin Lee, Howoong Jun, Jaeseok Heo, Jaewon Lee, Jeongwoo Oh, Songhwai Oh.

Figure 1
Figure 1. Figure 1: Overview of the RVN-Bench. The benchmark is designed for indoor mobile robots and focuses on collision-aware visual navigation, where an agent must reach sequential goal positions using only visual observations while avoiding collisions. Built upon Habitat 2.0 and utilizing HM3D scenes, it provides high-quality visual observations, an RL environment for training and evaluation, and pipelines for collecting… view at source ↗
Figure 2
Figure 2. Figure 2: Trajectory image dataset generation process. We generate trajectory image datasets through the following steps. (a) Start and goal positions are randomly sampled in the scene. (b) The shortest path between these positions is computed using the ground truth occupancy map padded with a margin larger than the agent’s radius. (c) The agent follows this path using discrete actions. (d) If the agent reaches the … view at source ↗
Figure 3
Figure 3. Figure 3: Overview of the NoMaD-Neg framework. (a) Two NoMaD￾PointGoal model, NoMaDe and NoMaDn, are trained separately with expert and negative trajectory datasets. (b) NoMaDe and NoMaDn predict expert trajectories ae,i ∈ aE and negative trajectories an,i ∈ aN , respectively. The expert trajectory a ∗ e with minimal CoR(ae,i, aE, aN ) is then selected. complex for goal encoding. We therefore replace the goal encode… view at source ↗
Figure 4
Figure 4. Figure 4: Real-world evaluation of NoMaD-PointGoal with different training datasets. (a) Clearpath Jackal UGV platform [32]–[34] used in the experiments. (b) Illustration of evaluation results in a house environment. Shown are the executed trajectories of the NoMaD-PointGoal agent trained on the real-world dataset (pink), trained on the simulation dataset (yellow), and trained on the combined real-world and simulati… 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

34 extracted references · 3 linked inside Pith

  1. [1]

    GNM: A general navigation model to drive any robot,

    D. Shah, A. Sridhar, A. Bhorkar, N. Hirose, and S. Levine, “GNM: A general navigation model to drive any robot,” inProc. of the IEEE International Conference on Robotics and Automation (ICRA), 2023

    2023

  2. [2]

    ViNT: A foundation model for visual navigation,

    D. Shah, A. Sridhar, N. Dashora, K. Stachowicz, K. Black, N. Hirose, and S. Levine, “ViNT: A foundation model for visual navigation,” in Proc. of the Conference on robot learning (CoRL), 2023

    2023

  3. [3]

    NoMaD: Goal masked diffusion policies for navigation and exploration,

    A. Sridhar, D. Shah, C. Glossop, and S. Levine, “NoMaD: Goal masked diffusion policies for navigation and exploration,” inProc. of the IEEE International Conference on Robotics and Automation (ICRA), 2024

    2024

  4. [4]

    Citywalker: Learning embodied urban navi- gation from web-scale videos,

    X. Liu, J. Li, Y . Jiang, N. Sujay, Z. Yang, J. Zhang, J. Abanes, J. Zhang, and C. Feng, “Citywalker: Learning embodied urban navi- gation from web-scale videos,” inProc. of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2025

    2025

  5. [5]

    CARLA: An open urban driving simulator,

    A. Dosovitskiy, G. Ros, F. Codevilla, A. Lopez, and V . Koltun, “CARLA: An open urban driving simulator,” inProc. of the Con- ference on robot learning (CoRL), 2017

    2017

  6. [6]

    SUMMIT: A simulator for urban driving in massive mixed traffic,

    P. Cai, Y . Lee, Y . Luo, and D. Hsu, “SUMMIT: A simulator for urban driving in massive mixed traffic,” inProc. of the IEEE International Conference on Robotics and Automation (ICRA), 2020

    2020

  7. [7]

    Habitat: A platform for embodied ai research,

    M. Savva, A. Kadian, O. Maksymets, Y . Zhao, E. Wijmans, B. Jain, J. Straub, J. Liu, V . Koltun, J. Maliket al., “Habitat: A platform for embodied ai research,” inProc. of the IEEE/CVF International Conference on Computer Vision (ICCV), 2019

    2019

  8. [8]

    Habitat 2.0: Training home assistants to rearrange their habitat,

    A. Szot, A. Clegg, E. Undersander, E. Wijmans, Y . Zhao, J. Turner, N. Maestre, M. Mukadam, D. S. Chaplot, O. Maksymetset al., “Habitat 2.0: Training home assistants to rearrange their habitat,” Advances in neural information processing systems (NeurIPS), 2021

    2021

  9. [9]

    Design and use paradigms for gazebo, an open-source multi-robot simulator,

    N. Koenig and A. Howard, “Design and use paradigms for gazebo, an open-source multi-robot simulator,” inProc. of the IEEE International Conference on Intelligent Robot Systems (IROS), 2004

    2004

  10. [10]

    Are we making real progress in simulated environments? measuring the sim2real gap in embodied visual navigation,

    Abhishek Kadian*, Joanne Truong*, A. Gokaslan, A. Clegg, E. Wij- mans, S. Lee, M. Savva, S. Chernova, and D. Batra, “Are we making real progress in simulated environments? measuring the sim2real gap in embodied visual navigation,” inarXiv:1912.06321, 2019

    Pith/arXiv arXiv 1912

  11. [11]

    HabiCrowd: A high performance simulator for crowd- aware visual navigation,

    A. Vuong, T. Nguyen, M. N. Vu, B. Huang, H. Binh, T. V o, and A. Nguyen, “HabiCrowd: A high performance simulator for crowd- aware visual navigation,” inProc. of the IEEE International Confer- ence on Intelligent Robot Systems (IROS), 2024

    2024

  12. [12]

    GOAT- bench: A benchmark for multi-modal lifelong navigation,

    M. Khanna, R. Ramrakhya, G. Chhablani, S. Yenamandra, T. Gervet, M. Chang, Z. Kira, D. S. Chaplot, D. Batra, and R. Mottaghi, “GOAT- bench: A benchmark for multi-modal lifelong navigation,” inProc. of the IEEE/CVF International Conference on Computer Vision (ICCV), 2024

    2024

  13. [13]

    Vision-and-language nav- igation: Interpreting visually-grounded navigation instructions in real environments,

    P. Anderson, Q. Wu, D. Teney, J. Bruce, M. Johnson, N. S ¨underhauf, I. Reid, S. Gould, and A. Van Den Hengel, “Vision-and-language nav- igation: Interpreting visually-grounded navigation instructions in real environments,” inProc. of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2018

    2018

  14. [14]

    Metaurban: A simulation platform for embodied ai in urban spaces,

    W. Wu, H. He, Y . Wang, C. Duan, J. He, Z. Liu, Q. Li, and B. Zhou, “Metaurban: A simulation platform for embodied ai in urban spaces,” International Conference on Learning Representation (ICLR), 2025

    2025

  15. [15]

    Habitat-matterport 3D dataset (hm3d): 1000 large-scale 3D environments for embodied ai,

    S. K. Ramakrishnan, A. Gokaslan, E. Wijmans, O. Maksymets, A. Clegg, J. Turner, E. Undersander, W. Galuba, A. Westbury, A. Chang, M. Savva, Y . Zhao, and D. Batra, “Habitat-matterport 3D dataset (hm3d): 1000 large-scale 3D environments for embodied ai,” in Proc. of the Neural Information Processing Systems (NeurIPS) Track on Datasets and Benchmarks, 2021

    2021

  16. [16]

    DD-PPO: Learning near-perfect pointgoal navigators from 2.5 billion frames,

    E. Wijmans, A. Kadian, A. Morcos, S. Lee, I. Essa, D. Parikh, M. Savva, and D. Batra, “DD-PPO: Learning near-perfect pointgoal navigators from 2.5 billion frames,” inInternational Conference on Learning Representations (ICLR), 2020

    2020

  17. [17]

    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

  18. [18]

    Benchmarking safe exploration in deep reinforcement learning,

    A. Ray, J. Achiam, and D. Amodei, “Benchmarking safe exploration in deep reinforcement learning,”arXiv preprint arXiv:1910.01708, 2019

    Pith/arXiv arXiv 1910

  19. [19]

    Habitat challenge 2023,

    K. Yadav, J. Krantz, R. Ramrakhya, S. K. Ramakrishnan, J. Yang, A. Wang, J. Turner, A. Gokaslan, V .-P. Berges, R. Mootaghi, O. Maksymets, A. X. Chang, M. Savva, A. Clegg, D. S. Chaplot, and D. Batra, “Habitat challenge 2023,” https://aihabitat.org/challenge/ 2023/, 2023

    2023

  20. [20]

    HM3D- OVON: A dataset and benchmark for open-vocabulary object goal navigation,

    N. Yokoyama, R. Ramrakhya, A. Das, D. Batra, and S. Ha, “HM3D- OVON: A dataset and benchmark for open-vocabulary object goal navigation,” inProc. of the IEEE International Conference on Intelli- gent Robot Systems (IROS), 2024

    2024

  21. [21]

    Autonomous ground navigation in highly constrained spaces: Lessons learned from the benchmark autonomous robot navigation challenge at icra 2022 competitions,

    X. Xiao, Z. Xu, Z. Wang, Y . Song, G. Warnell, P. Stone, T. Zhang, S. Ravi, G. Wang, H. Karnanet al., “Autonomous ground navigation in highly constrained spaces: Lessons learned from the benchmark autonomous robot navigation challenge at icra 2022 competitions,” IEEE Robotics and Automation Magazine, 2022

    2022

  22. [22]

    Vision and navigation for the carnegie-mellon navlab,

    C. Thorpe, M. H. Hebert, T. Kanade, and S. A. Shafer, “Vision and navigation for the carnegie-mellon navlab,”IEEE Transactions on Pattern Analysis and Machine Intelligence, 1988

    1988

  23. [23]

    Learning by cheating,

    D. Chen, B. Zhou, V . Koltun, and P. Kr ¨ahenb¨uhl, “Learning by cheating,” inProc. of the Conference on robot learning (CoRL), 2020

    2020

  24. [24]

    Autonomous quadrotor navigation with vision based obstacle avoidance and path planning,

    H.-Y . Lin and X.-Z. Peng, “Autonomous quadrotor navigation with vision based obstacle avoidance and path planning,”IEEE Access, 2021

    2021

  25. [25]

    A formal basis for the heuristic determination of minimum cost paths,

    P. E. Hart, N. J. Nilsson, and B. Raphael, “A formal basis for the heuristic determination of minimum cost paths,”IEEE transactions on Systems Science and Cybernetics, 1968

    1968

  26. [26]

    Depth anything v2,

    L. Yang, B. Kang, Z. Huang, Z. Zhao, X. Xu, J. Feng, and H. Zhao, “Depth anything v2,”Advances in neural information processing systems (NeurIPS), 2024

    2024

  27. [27]

    Mixgail: Autonomous driving using demonstrations with mixed qualities,

    G. Lee, D. Kim, W. Oh, K. Lee, and S. Oh, “Mixgail: Autonomous driving using demonstrations with mixed qualities,” inProc. of the IEEE International Conference on Intelligent Robot Systems (IROS), 2020

    2020

  28. [28]

    Deep visual mpc-policy learning for navigation,

    N. Hirose, F. Xia, R. Mart ´ın-Mart´ın, A. Sadeghian, and S. Savarese, “Deep visual mpc-policy learning for navigation,”IEEE Robotics and Automation Letters (RA-L), 2019

    2019

  29. [29]

    Rapid explo- ration for open-world navigation with latent goal models,

    D. Shah, B. Eysenbach, N. Rhinehart, and S. Levine, “Rapid explo- ration for open-world navigation with latent goal models,” inProc. of the Conference on robot learning (CoRL), 2021

    2021

  30. [30]

    Socially compliant navigation dataset (scand): A large-scale dataset of demonstrations for social navigation,

    H. Karnan, A. Nair, X. Xiao, G. Warnell, S. Pirk, A. Toshev, J. Hart, J. Biswas, and P. Stone, “Socially compliant navigation dataset (scand): A large-scale dataset of demonstrations for social navigation,”IEEE Robotics and Automation Letters (RA-L), 2022

    2022

  31. [31]

    Sacson: Scalable autonomous control for social navigation,

    N. Hirose, D. Shah, A. Sridhar, and S. Levine, “Sacson: Scalable autonomous control for social navigation,”IEEE Robotics and Au- tomation Letters (RA-L), 2023

    2023

  32. [32]

    Jackal UGV - Small Weatherproof Robot,

    Clearpath Robotics, “Jackal UGV - Small Weatherproof Robot,” https: //clearpathrobotics.com/jackal-small-unmanned-ground-vehicle/, Dec. 2020, accessed: 2025-09-01

    2020

  33. [33]

    ZED-X Stereo Camera – Rugged AI Stereo Camera for Robotics,

    Stereolabs, “ZED-X Stereo Camera – Rugged AI Stereo Camera for Robotics,” https://www.stereolabs.com/products/zed-x, Dec. 2022, accessed: 2025-09-01

    2022

  34. [34]

    NVIDIA Jetson Orin – AI Computing Modules and Devel- oper Kits for Robotics and Edge AI,

    NVIDIA, “NVIDIA Jetson Orin – AI Computing Modules and Devel- oper Kits for Robotics and Edge AI,” https://www.nvidia.com/en-us/ autonomous-machines/embedded-systems/jetson-orin/, Sep. 2025, ac- cessed: 2025-09-01

    2025