Pith. sign in

REVIEW 3 major objections 4 minor 80 references

Force-Based Robotic Imitation Learning: A Two-Phase Approach for Construction Assembly Tasks

T0 review · 3 major / 4 minor · reviewed 2026-08-10 · deepseek-v4-flash

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

Pith's one-line read Force feedback from human demonstrations, distilled through GAIL and PPO, beats vision and pure RL for simulated pipe insertion.

desk verdict A real haptic demonstration pipeline for construction assembly, but the force-vs-visual comparison is confounded by a reward only the force group can perceive. read the letter →

arxiv 2501.14942 v1 pith:V2ZO6ZV7 submitted 2025-01-24 cs.RO cs.AI

Hengxu You , Yang Ye , Tianyu Zhou , Jing Du This is my paper · ORCID
classification cs.ROcs.AI
keywords imitationlearningforcefeedbackhapticdemonstrationconstructionroboticspipeinsertiongenerativeadversarialproximalpolicyoptimizationdigitaltwin
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

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

The reading

The paper sets out to show that force feedback, not vision, should carry imitation learning for contact-rich construction tasks such as pipe insertion, where the insertion point is occluded and surface irregularities defeat visual cues. Its two-phase framework first records human demonstrations on a robot arm whose Unity digital twin generates haptic collision feedback, so the demonstrations themselves contain human force responses, and then converts those demonstrations into robot policy via Generative Adversarial Imitation Learning (GAIL) initialization refined by Proximal Policy Optimization (PPO), with normal and friction forces folded into the reward. The reported result is that force-trained policies complete simulated pipe insertion faster and more reliably: average success of 81.2% and 74.1% under two randomization conditions, against 27.6% and 25.7% for visual demonstration training and 24% and 14% for pure reinforcement learning. If the result holds beyond simulation, construction robots could learn dexterous assembly from human touch rather than from cameras.

What carries the argument

The mechanism that carries the argument is the two-phase DCI-LG loop with its custom contact-force core. In the Data Collection through Interaction (DCI) phase, a human pushes a pipe-handle mounted on a 6-DoF Panda arm while ROS-Sharp streams the pose into a Unity digital twin; contact is detected with trigger volumes instead of the full physics solver, and the impedance controller $F = \delta_{coll}\delta_{dir}(M_1\ddot{x}_{dist} + B_1\dot{x}_{dist} + K_1 x_{dist}) + g_c$ with a very large stiffness $K_1$ turns virtual collisions into haptic resistance on the real arm. The direction filter $\delta_{dir}$ is the piece that makes the force feel physical: resistance is applied only while the operator is pushing toward the last collision point, and it vanishes when the operator pulls away. Collision impulses are scaled by a predefined pipe mass $m$ and frame-to-frame velocity, the contact normal comes from a raycast fan around the closest contact point whose covariance is decomposed by singular value decomposition, and the resulting normal force $\mathbf{F}_{normal}$ and friction $\mathbf{F}_{friction}$ form the force observation $\mathbf{F}_{obs}$. In the Learning with Guidance (LG) phase, this observation plus depth and distance enters a GAIL generator with a 256-neuron hidden layer that emits a 3-D Cartesian force action, and the policy is then fine-tuned with PPO; the reward $r_t$ pays positive score when both normal and friction forces shrink across steps and adds +1 when the pipe reaches the 0.5 m target depth, which is what presses the agent toward insertion through decreasing contact rather than brute force.

What would settle it

Replay the recorded human motions on a physical pipe-insertion rig instrumented with a force-torque sensor and compare the measured normal and friction profiles with what the Unity simulation outputs for the same trajectories; then train the GAIL-PPO pipeline on real-force demonstrations and run the learned policy on the physical robot. If the simulated contact forces differ substantially from measured ones, or if policies trained on real forces do not approach the reported 74-81% success rates, then the force-modality advantage is a property of the simulator, not of force feedback as such.

Watch

Extended reading notes

Core claim

The paper's central claim is that force feedback is a better demonstration modality than vision for training contact-rich construction assembly, and that the quality of force demonstrations depends on letting the human operator feel realistic contact while collecting them. The authors implement this claim as a two-phase system: humans first guide a physical robot arm through pipe insertion while a Unity-simulated twin reflects collisions back as haptic resistance, so the 20 recorded demonstrations per condition encode real human correction behavior under contact; then Generative Adversarial Imitation Learning (GAIL) converts those demonstrations into an initial policy, which Proximal Policy Optimization (PPO) refines, with normal and friction forces embedded in the observation and reward. On inference tests with randomized pipe positions, force-trained agents succeed at an average of 81.2% over 100 trials per demonstration versus 27.6% for visually trained agents, and the gap persists (74.1% versus 25.7%) when the target location is also randomized; pure reinforcement learning without force observations reaches only 24% and 14% in the two conditions. The authors further report that force demonstrations converge faster in training and produce narrower reward distributions, though they acknowledge two outlier demonstrations that underperform their visual counterparts.

Load-bearing premise

The whole framework rests on the assumption that the Unity simulation's contact forces, including stiffness resistance, collision impulses, and computed surface normals, feel and behave enough like real pipe-insertion forces that human adjustments collected in the rig and the policies trained on them transfer to a physical robot, and the paper reports no real force measurements to validate that.

Editorial extensions

If this is right

  • Force-guided trials converge faster than visual-guided trials for all 20 demonstrations per group, so training time is cut when force demonstrations are used.
  • Force-trained policies keep most of their edge under randomization: 74.1% average success when both pipe positions are randomized, versus 25.7% for visual policies, indicating the force observation carries transferable alignment information.
  • Imitation initialization adds to the force feedback: the pure-RL condition (position-only observation) converges to roughly 70% of the force-feedback reward, and every force demonstration trained with the DIL-DRL workflow outperforms the same force observations trained with DRL alone, so both the force modality and the imitation step contribute.
  • Force demonstrations produce narrower performance distributions, meaning more consistent policies, though two of the twenty force demonstrations (trials 8 and 20) converge slower and generalize worse than their visual counterparts.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • Because the paper never measures real contact forces, the decisive extension it implies but does not run is a sim-to-real check: replay recorded human motions on an instrumented physical insertion rig and compare measured normal and friction profiles with the simulated ones; the reward function's assumptions stand or fall on that match.
  • The reward design, paying positive score for shrinking normal and friction forces, is stated only for pipe insertion, but the same shaping should apply to any peg-in-hole or alignment-through-contact task, so the claimed benefit is plausibly a general recipe for contact-rich assembly rather than a pipe-specific trick; the paper does not test that generality.
  • The two outlier demonstrations suggest a practical selection rule the paper only gestures at: filter or reweight demonstrations by force-profile consistency before GAIL initialization, which could remove the policies that drag the average down.
  • A data-efficiency argument is implicit in the architecture: the force observation is eight numbers versus 258 for the visual encoder, so if the success-rate gap holds, force feedback is also the cheaper demonstration modality; quantifying the sample-complexity difference would make that case explicit.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

3 major / 4 minor

Summary. The paper proposes a two-phase system for learning pipe-insertion assembly: in the first phase, a human operator guides a Franka Emika arm that is synchronized with a Unity simulator, receiving haptic feedback from a custom collision/force model; in the second phase, the recorded demonstrations are used to initialize a policy via GAIL, which is then fine-tuned with PPO. The authors compare agents trained on force observations (F_obs) against agents trained on ray-caster visual observations (Vis_t) and against reward-only RL baselines, reporting faster convergence, higher success rates, and shorter episode lengths for the force-based agents across 20 demonstrations per condition. The abstract and conclusion claim that force feedback is a superior demonstration modality for contact-rich construction assembly tasks.

Significance. If the central comparison were valid, the paper would provide a useful practical contribution: an end-to-end human-in-the-loop data-collection and learning pipeline that uses force feedback for contact-rich assembly, with a moderately large set of demonstrations and 100 inference trials per learned policy. The authors also explicitly examine within-group variability in Section 5.4, which is a commendable check on the robustness of their results. However, the headline claim is currently undermined by a methodological confound: the reward in Eq. (11) directly rewards reductions in the very quantities that appear only in the force observation, so the reported advantage may reflect reward-observation alignment rather than any benefit of force demonstrations as a modality. The custom Unity force model is also unvalidated against physical measurements, which limits the external validity of any simulation-only conclusion. With the confounds removed or controlled and with statistical inference added, the question would be well worth revisiting.

major comments (3)
  1. [Section 3.4.2, Eq. (11)] The central comparison is confounded. The reward r_t in Eq. (11) gives positive feedback exactly when ||F_normal^t|| and ||F_friction^t|| decrease relative to the previous step, and the force observation F_obs is defined as [F_normal, F_friction]. The visual observation Vis_t = {d, l, [P]} contains no force channel, and the RL baseline observes neither forces nor ray-caster points. Consequently, the force-condition agent is rewarded for reducing quantities that are directly in its state, while the visual and RL conditions are rewarded for reducing quantities they cannot perceive. This means the experiments vary at least two things at once: demonstration modality (force vs. visual) and alignment between the reward and the agent's observation. The paper's conclusion in Sections 5 and 6 that force demonstrations outperform visual demonstrations therefore does not follow from the reported results. I would need an ablation that fixes the reward-observation alignment, for example by giving the visual condition the same force-decrease reward computed from the simulator's hidden state, or by removing the force-decrease component from the force condition's reward, or by training a force-observation agent on visual/hand-crafted demonstrations.
  2. [Section 3.4 and Section 6] The external validity of the force model is unsupported. The simulator's contact forces are produced by a custom impedance-control and impulse model whose parameters (mass m, stiffness K_d, ray count n_rays, 45-degree threshold) are chosen without calibration against any physical force-torque measurement. The paper states in Section 6 that validation with real-world data is future work, but the abstract and introduction describe the simulation as providing 'realistic' or 'similar-to-real' force interactions. Since the demonstrations, the agent observations, the reward, and the success criterion all come from this same unvalidated model, a reader cannot determine whether the learned force-based policies would transfer to a physical pipe-insertion task. I recommend either calibrating the model against measured contact forces or substantially tempering the real-world claims and explicitly labeling the results as simulation-only.
  3. [Section 5.2, 5.3, and 5.4, Tables 2-5] The superiority claim is reported without any statistical inference. Each trial yields a success rate and mean episode length over 100 inferences, but the paper only reports means and standard deviations across the 20 demonstrations and does not test whether the force-group distribution differs from the visual-group distribution. This matters because Section 5.4 identifies two force demonstrations (8 and 20) whose inference success rates are below their visual counterparts; with 20 paired samples and this level of variability, the headline difference could be driven by a few large successes. I would like to see paired per-demonstration differences with a nonparametric test (e.g., Wilcoxon signed-rank), confidence intervals for success rates, and an explicit statement about how many of the 20 force demonstrations beat their visual counterpart.
minor comments (4)
  1. [Section 5.2, Tables 3 and 4] The visual-group table for randomized inner pipe location is labeled 'Table 4' but should be Table 3, and the cross-references in Sections 5.2 and 5.3 are inconsistent; please renumber all tables and fix the in-text references.
  2. [Section 3.3, Eq. (1)] In Eq. (1), K_d is described as a 'damping coefficient for motion speed,' but it appears as the coefficient of the displacement term in an impedance control law and should be labeled as a stiffness coefficient; please clarify the physical meaning and units of M_d, B_d, and K_d.
  3. [Section 3.4.2, Eq. (10)] The sign convention for the indicator Ind in Eq. (10) appears inconsistent with the earlier definition of delta_move in Eq. (2): if Ind >= 0, the paper sets delta_c to 0, yet a positive dot product would normally mean the handler is still pushing toward the collision point; please reconcile the sign convention and the verbal description.
  4. [Section 1 and Section 4] The paper uses promotional language such as 'groundbreaking' and contains typographical errors such as 'EXPEERIMENTS'; I recommend a careful language edit and neutral terminology throughout.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: the paper's force-versus-visual comparison is empirical, and the force-group advantage is not entailed by construction.

full rationale

The paper does not derive a prediction from first principles; it reports an empirical reinforcement-learning comparison. The force observation in Eq. (8) and the reward in Eq. (11) both contain the normal and friction force magnitudes, which creates a reward-observation alignment confound, but this is an experimental design issue rather than a circular reduction: the claim that force demonstrations outperform visual demonstrations is not equivalent to the reward definition, and Section 5.4 shows two force demonstrations underperformed their visual counterparts, so the outcome was not forced by construction. The custom Unity force model is an external-validity limitation acknowledged in Section 6, not a circular input. Self-citations (refs 9, 10, 11, 19, 37, 40) are background citations and are not load-bearing; no uniqueness theorem or ansatz is imported from them. The evaluation is self-contained against internal baselines (RL, visual, and force), and no fitted parameter is renamed as a prediction.

Assumptions & free parameters 6 free parameters · 6 assumptions · 0 invented entities

No new physical entities are postulated; the system is a software integration of existing robots, simulators, and learning algorithms. The central claim rests on hand-set simulation parameters, an unvalidated force model, and a hand-designed reward function, all of which are internal to the paper.

free parameters (6)
  • Impedance control coefficients M_d, B_d, K_d = Not reported
    Eq (1) defines the simulated resistance force with inertia, damping, and stiffness terms; K_d is set very large during collision, but no numeric values or sensitivity analysis are given, and the force feedback quality depends directly on them.
  • Inner pipe mass m = Not reported
    Used in Eq (3) to compute impulse force magnitude F_imp = m * |v_t| / T; the simulated contact forces sent to the human depend on this hand-set value.
  • Ray count n_rays and 45-degree angle threshold = Not reported (n_rays unspecified; angle 45 degrees)
    These parameters control the ray-caster used to estimate the surface normal via SVD in Section 3.4.2; no sensitivity analysis is provided.
  • Reward normalization 1/MaxStep and force-decrease reward weights = 1/MaxStep; +1, 0, -1 per step
    Eq (11) is hand-designed to reward decreasing normal and friction forces; the weights and the threshold for equal forces are choices not derived from task specifications.
  • GAIL and behavioral cloning strengths = GAIL 0.01, BC 1.0
    These hyperparameters in Section 4 control how much the demonstrations shape the policy; their values are chosen without ablation.
  • Success depth 0.5m and pad sensor = 0.5m
    The task completion threshold is set by the authors; results depend on this definition, and it is not varied in experiments.
assumptions (6)
  • domain assumption Rigid-body assumption: the metal pipe handle and pipes have negligible deformation under operational loads, so impedance control can represent contact forces.
    Invoked in Section 3.3 when choosing the handle and in Section 3.4 when treating pipes as rigid bodies; real pipes deform slightly, and the force model ignores deformation.
  • domain assumption Unity PhysX OnTrigger detection plus the custom impulse and SVD force model faithfully represents pipe insertion contact dynamics.
    Section 3.4 replaces PhysX's physics solver with trigger-based detection and a custom mathematical model; no real force-torque measurements validate this proxy.
  • ad hoc to paper The reward function in Eq (11) is an appropriate objective for learning pipe insertion.
    The reward explicitly rewards decreasing normal and friction forces and completion depth; this embeds the paper's force-based control preference into the training objective, making the force group's advantage partly by construction.
  • domain assumption Demonstrations collected in VR and Unity transfer to useful policies for the simulated robot and, by implication, to real robots.
    The authors state future work will validate with real-world data (Section 6), acknowledging that this transfer is not demonstrated.
  • standard math SVD of sampled raycast points yields the correct surface normal for impulse direction.
    Section 3.4.2 uses the least significant singular vector of A as the surface normal; this is a standard geometric estimation approach.
  • standard math Newton's second law and impulse-momentum relation apply at the single-frame collision timescale.
    Eq (3) computes impulse as m*|v|/T; this is standard mechanics but ignores restitution and multi-frame contact.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Force-Based Robotic Imitation Learning: A Two-Phase Approach for Construction Assembly Tasks." pith.science (2026). https://pith.science/paper/V2ZO6ZV7

@misc{pith2026250114942,
  author       = {Pith},
  title        = {Pith review of: Force-Based Robotic Imitation Learning: A Two-Phase Approach for Construction Assembly Tasks},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/V2ZO6ZV7}},
  note         = {Machine review of arXiv:2501.14942}
}
read the original abstract

The drive for efficiency and safety in construction has boosted the role of robotics and automation. However, complex tasks like welding and pipe insertion pose challenges due to their need for precise adaptive force control, which complicates robotic training. This paper proposes a two-phase system to improve robot learning, integrating human-derived force feedback. The first phase captures real-time data from operators using a robot arm linked with a virtual simulator via ROS-Sharp. In the second phase, this feedback is converted into robotic motion instructions, using a generative approach to incorporate force feedback into the learning process. This method's effectiveness is demonstrated through improved task completion times and success rates. The framework simulates realistic force-based interactions, enhancing the training data's quality for precise robotic manipulation in construction tasks.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

80 extracted references · 75 canonical work pages

  1. [1]

    1 Force-Based Robotic Imitation Learning: A Two-Phase Approach for Construction Assembly Tasks Hengxu You1, Yang Ye, Ph.D.2, Tianyu Zhou, Ph.D.3, Eric Jing Du, Ph.D.4* 1 Ph.D. Student, Engineering School of Sustainable Infrastructure & Environment, University of Florida, 1949 Stadium Road 454A Weil Hall, Gainesville, FL 32611; email: you.h@ufl.edu 2 Assis...

    work page 1949

  2. [2]

    Test results of inner pipe randomization for force group (100 trials each). Trial RL Base 1 2 3 4 5 6 7 8 9 10 Success 24 40 100 80 94 89 75 79 85 41 83 80 Mean 3500 2149 458 918 973 46 1015 994 783 2119 848 794 STD 55.20 40.23 9.83 21.45 34.06 5.04 32.96 19.95 14.08 95.85 23.78 12.43 Trial 11 12 13 14 15 16 17 18 19 20 Success 77` 95 85 76 88 79 90 84 92...

    work page 1930

  3. [3]

    Hyperparameters for Training Training Settings Network Setting trainer PPO hidden units 128 batch size 128 number of hidden 2 buffer size 2048 visual encoder Conv_2d number of epochs 3 Update Setting learning rate 0.0003 total steps 5,000,000 regularization 0.95 max steps 10,000 reward discount 0.99 summary steps 12,000 To train the force and visual group...

    work page

  4. [4]

    EXPEERIMENTS The physical interaction simulation system is designed in Unity using the physical engine. In our experiments, the outer pipe exists only in Unity and its position is directly obtained, while the digital twin is built for the robot arm to collect human expert action data through VR. The training process is built with ML-Agents which is a well...

    work page 2018

  5. [5]

    Test results of target randomization for force group (100 trials each). Trial RL Base 1 2 3 4 5 6 7 8 9 10 Success 14 46 74 76 76 88 80 70 84 35 72 72 Mean 2141 1882 905 995 1063 848 871 1140 699 1914 945 932 STD 32.98 33.49 14.98 19.62 17.12 16.34 14.25 23.02 11.96 49.73 13.84 17.02 Trial 11 12 13 14 15 16 17 18 19 20 Success 67 98 72 73 86 77 76 94 76 3...

    work page 1914

  6. [6]

    Abioye, L.O

    S.O. Abioye, L.O. Oyedele, L. Akanbi, A. Ajayi, J.M.D. Delgado, M. Bilal, O.O. Akinade, A. Ahmed, Artificial intelligence in the construction industry: A review of present status, opportunities and future challenges, Journal of Building Engineering, 44 (2021) 103299

    work page 2021

  7. [7]

    Baduge, S

    S.K. Baduge, S. Thilakarathna, J.S. Perera, M. Arashpour, P. Sharafi, B. Teodosio, A. Shringi, P. Mendis, Artificial intelligence and smart vision for building and construction 4.0: Machine and deep learning methods and applications, Automation in Construction, 141 (2022) 104440

    work page 2022

  8. [8]

    Pradhananga, M

    P. Pradhananga, M. ElZomor, G. Santi Kasabdji, Identifying the challenges to adopting robotics in the US construction industry, Journal of Construction Engineering and Management, 147 (2021) 05021003

    work page 2021

Show all 80 references
  1. [9]

    Halder, K

    S. Halder, K. Afsari, Robots in inspection and monitoring of buildings and infrastructure: A systematic review, Applied Sciences, 13 (2023)

    2023

  2. [10]

    Sudhakaran, K

    S. Sudhakaran, K. Montgomery, M. Kashef, D. Cavalcanti, R. Candell, Wireless time sensitive networking for industrial collaborative robotic workcells, 2021 17th IEEE International Conference on Factory Communication Systems (WFCS), IEEE, 2021, pp. 91-94

    2021

  3. [11]

    Surati, S

    S. Surati, S. Hedaoo, T. Rotti, V . Ahuja, N. Patel, Pick and place robotic arm: a review paper, Int. Res. J. Eng. Technol, 8 (2021) 2121-2129

    2021

  4. [12]

    Kim, J.-M

    U.-H. Kim, J.-M. Park, T.-J. Song, J.-H. Kim, 3-D scene graph: A sparse and semantic representation of physical environments for intelligent agents, IEEE transactions on cybernetics, 50 (2019) 4921-4933

    2019

  5. [13]

    Matulis, C

    M. Matulis, C. Harvey, A robot arm digital twin utilising reinforcement learning, Computers & Graphics, 95 (2021) 106-114

    2021

  6. [14]

    Y . Ye, T. Zhou, J. Du, Robot-assisted immersive kinematic experience transfer for welding training, Journal of Computing in Civil Engineering, 37 (2023) 04023002

    2023

  7. [15]

    Z.-H. Yin, W. Ye, Q. Chen, Y . Gao, Planning for Sample Efficient Imitation Learning, Advances in Neural Information Processing Systems, 35 (2022) 2577-2589

    2022

  8. [16]

    Hussein, M.M

    A. Hussein, M.M. Gaber, E. Elyan, C. Jayne, Imitation learning: A survey of learning methods, ACM Computing Surveys (CSUR), 50 (2017) 1-35

    2017

  9. [17]

    Yu, V .R

    H. Yu, V .R. Kamat, C.C. Menassa, Cloud-Based Hierarchical Imitation Learning for Scalable Transfer of Construction Skills from Human Workers to Assisting Robots, Journal of Computing in Civil Engineering, 38 (2024) 04024019

    2024

  10. [18]

    J. Ho, S. Ermon, Generative adversarial imitation learning, Advances in neural information processing 33 systems, 29 (2016)

    2016

  11. [19]

    Torabi, G

    F. Torabi, G. Warnell, P. Stone, Recent advances in imitation learning from observation, arXiv preprint arXiv:1905.13566, (2019)

    2019 arXiv

  12. [20]

    C. Finn, T. Yu, T. Zhang, P. Abbeel, S. Levine, One-shot visual imitation learning via meta-learning, Conference on robot learning, PMLR, 2017, pp. 357-368

    2017

  13. [21]

    Z. Xia, Z. Deng, B. Fang, Y . Yang, F. Sun, A review on sensory perception for dexterous robotic manipulation, International Journal of Advanced Robotic Systems, 19 (2022) 17298806221095974

    2022

  14. [22]

    Q. Zhu, T. Zhou, J. Du, Haptics-based force balance controller for tower crane payload sway controls, Automation in Construction, 144 (2022) 104597

    2022

  15. [23]

    Valdivia, R

    A.A. Valdivia, R. Shailly, N. Seth, F. Fuentes, D.P. Losey, L.H. Blumenschein, Wrapped haptic display for communicating physical robot learning, 2022 IEEE 5th International Conference on Soft Robotics (RoboSoft), IEEE, 2022, pp. 823-830

    2022

  16. [24]

    El Rassi, J.-M

    I. El Rassi, J.-M. El Rassi, A review of haptic feedback in tele-operated robotic surgery, Journal of medical engineering & technology, 44 (2020) 247-254

    2020

  17. [25]

    L. Chen, H. Luo, Near-optimal goal-oriented reinforcement learning in non-stationary environments, Advances in Neural Information Processing Systems, 35 (2022) 33973-33984

    2022

  18. [26]

    Singh, R

    B. Singh, R. Kumar, V .P. Singh, Reinforcement learning in robotic applications: a comprehensive survey, Artificial Intelligence Review, (2022) 1-46

    2022

  19. [27]

    Zheng, S

    B. Zheng, S. Verma, J. Zhou, I. Tsang, F. Chen, Imitation learning: Progress, taxonomies and challenges, arXiv preprint arXiv:2106.12177, (2021)

    2021 arXiv

  20. [28]

    Arulkumaran, D.O

    K. Arulkumaran, D.O. Lillrank, A pragmatic look at deep imitation learning, arXiv preprint arXiv:2108.01867, (2021)

    2021 arXiv

  21. [29]

    M. Liu, W. Buntine, G. Haffari, Learning how to actively learn: A deep imitation learning approach, Proceedings of the 56th Annual Meeting of the Association for Computational Linguistics (V olume 1: Long Papers), 2018, pp. 1874-1883

    2018

  22. [30]

    D. Lee, S. Lee, N. Masoud, M. Krishnan, V .C. Li, Digital twin-driven deep reinforcement learning for adaptive task allocation in robotic construction, Advanced Engineering Informatics, 53 (2022) 101710

    2022

  23. [31]

    K. Xu, Z. Hu, R. Doshi, A. Rovinsky, V . Kumar, A. Gupta, S. Levine, Dexterous manipulation from images: Autonomous real-world rl via substep guidance, 2023 IEEE International Conference on Robotics and Automation (ICRA), IEEE, 2023, pp. 5938-5945

    2023

  24. [32]

    J. Hua, L. Zeng, G. Li, Z. Ju, Learning for a robot: Deep reinforcement learning, imitation learning, transfer learning, Sensors, 21 (2021)

    2021

  25. [33]

    Ibarz, J

    J. Ibarz, J. Tan, C. Finn, M. Kalakrishnan, P. Pastor, S. Levine, How to train your robot with deep reinforcement learning: lessons we have learned, The International Journal of Robotics Research, 40 (2021) 698-721

    2021

  26. [34]

    Tarbouriech, M

    J. Tarbouriech, M. Pirotta, M. Valko, A. Lazaric, A provably efficient sample collection strategy for reinforcement learning, Advances in Neural Information Processing Systems, 34 (2021) 7611-7624

    2021

  27. [35]

    Wenzel, T

    P. Wenzel, T. Schön, L. Leal-Taixé, D. Cremers, Vision-based mobile robotics obstacle avoidance with deep reinforcement learning, 2021 IEEE International Conference on Robotics and Automation (ICRA), 34 IEEE, 2021, pp. 14360-14366

    2021

  28. [36]

    Everett, Y .F

    M. Everett, Y .F. Chen, J.P. How, Collision avoidance in pedestrian-rich environments with deep reinforcement learning, IEEE Access, 9 (2021) 10357-10377

    2021

  29. [37]

    G. Kahn, A. Villaflor, B. Ding, P. Abbeel, S. Levine, Self-supervised deep reinforcement learning with generalized computation graphs for robot navigation, 2018 IEEE international conference on robotics and automation (ICRA), IEEE, 2018, pp. 5129-5136

    2018

  30. [38]

    Pérez-D’Arpino, C

    C. Pérez-D’Arpino, C. Liu, P. Goebel, R. Martín-Martín, S. Savarese, Robot navigation in constrained pedestrian environments using reinforcement learning, 2021 IEEE International Conference on Robotics and Automation (ICRA), IEEE, 2021, pp. 1140-1146

    2021

  31. [39]

    T. Zhou, Q. Zhu, J. Du, Intuitive robot teleoperation for civil engineering operations with virtual reality and deep learning scene reconstruction, Advanced Engineering Informatics, 46 (2020) 101170

    2020

  32. [40]

    D. Liu, Z. Wang, B. Lu, M. Cong, H. Yu, Q. Zou, A reinforcement learning-based framework for robot manipulation skill acquisition, IEEE Access, 8 (2020) 108429-108437

    2020

  33. [41]

    El-Fakdi, M

    A. El-Fakdi, M. Carreras, Two-step gradient-based reinforcement learning for underwater robotics behavior learning, Robotics and Autonomous Systems, 61 (2013) 271-282

    2013

  34. [42]

    P. Xia, F. Xu, T. Zhou, J. Du, Benchmarking human versus robot performance in emergency structural inspection, Journal of Construction Engineering and Management, 148 (2022) 04022070

    2022

  35. [43]

    Niroui, K

    F. Niroui, K. Zhang, Z. Kashino, G. Nejat, Deep reinforcement learning robot for search and rescue applications: Exploration in unknown cluttered environments, IEEE Robotics and Automation Letters, 4 (2019) 610-617

    2019

  36. [44]

    J. Hao, T. Yang, H. Tang, C. Bai, J. Liu, Z. Meng, P. Liu, Z. Wang, Exploration in deep reinforcement learning: From single-agent to multiagent domain, IEEE Transactions on Neural Networks and Learning Systems, (2023)

    2023

  37. [45]

    Delgado, L

    J.M.D. Delgado, L. Oyedele, Robotics in construction: A critical review of the reinforcement learning and imitation learning paradigms, Advanced Engineering Informatics, 54 (2022) 101787

    2022

  38. [46]

    D.K. Jha, S. Jain, D. Romeres, W. Yerazunis, D. Nikovski, Generalizable human-robot collaborative assembly using imitation learning and force control, 2023 European Control Conference (ECC), IEEE, 2023, pp. 1-8

    2023

  39. [47]

    Sasagawa, K

    A. Sasagawa, K. Fujimoto, S. Sakaino, T. Tsuji, Imitation learning based on bilateral control for human–robot cooperation, IEEE Robotics and Automation Letters, 5 (2020) 6169-6176

    2020

  40. [48]

    Liang, V .R

    C.-J. Liang, V .R. Kamat, C.C. Menassa, Teaching robots to perform quasi-repetitive construction tasks through human demonstration, Automation in Construction, 120 (2020) 103370

    2020

  41. [49]

    Y . Lu, J. Fu, G. Tucker, X. Pan, E. Bronstein, B. Roelofs, B. Sapp, B. White, A. Faust, S. Whiteson, Imitation is not enough: Robustifying imitation with reinforcement learning for challenging driving scenarios, arXiv preprint arXiv:2212.11419, (2022)

    2022 arXiv

  42. [50]

    Fujimoto, S.S

    S. Fujimoto, S.S. Gu, A minimalist approach to offline reinforcement learning, Advances in neural information processing systems, 34 (2021) 20132-20145

    2021

  43. [51]

    Rajeswaran, V

    A. Rajeswaran, V . Kumar, A. Gupta, G. Vezzani, J. Schulman, E. Todorov, S. Levine, Learning complex 35 dexterous manipulation with deep reinforcement learning and demonstrations, arXiv preprint arXiv:1709.10087, (2017)

    2017 arXiv

  44. [52]

    Vecerik, T

    M. Vecerik, T. Hester, J. Scholz, F. Wang, O. Pietquin, B. Piot, N. Heess, T. Rothörl, T. Lampe, M. Riedmiller, Leveraging demonstrations for deep reinforcement learning on robotics problems with sparse rewards, arXiv preprint arXiv:1707.08817, (2017)

    2017 arXiv

  45. [53]

    Huang, Z

    L. Huang, Z. Zhu, Z. Zou, To imitate or not to imitate: Boosting reinforcement learning-based construction robotic control for long-horizon tasks using virtual demonstrations, Automation in Construction, 146 (2023) 104691

    2023

  46. [54]

    Lundeen, V .R

    K.M. Lundeen, V .R. Kamat, C.C. Menassa, W. McGee, Scene understanding for adaptive manipulation in robotized construction work, Automation in Construction, 82 (2017) 16-30

    2017

  47. [55]

    Lundeen, V .R

    K.M. Lundeen, V .R. Kamat, C.C. Menassa, W. McGee, Autonomous motion planning and task execution in geometrically adaptive robotized construction work, Automation in Construction, 100 (2019) 24-45

    2019

  48. [56]

    Wang, C.-J

    X. Wang, C.-J. Liang, C.C. Menassa, V .R. Kamat, Interactive and immersive process-level digital twin for collaborative human–robot construction work, Journal of Computing in Civil Engineering, 35 (2021) 04021023

    2021

  49. [57]

    Liang, X

    C.-J. Liang, X. Wang, V .R. Kamat, C.C. Menassa, Human–robot collaboration in construction: Classification and research trends, Journal of Construction Engineering and Management, 147 (2021) 03121006

    2021

  50. [58]

    J. Cai, A. Du, X. Liang, S. Li, Prediction-based path planning for safe and efficient human–robot collaboration in construction via deep reinforcement learning, Journal of Computing in Civil Engineering, 37 (2023) 04022046

    2023

  51. [59]

    Liang, K.M

    C.-J. Liang, K.M. Lundeen, W. McGee, C.C. Menassa, S. Lee, V .R. Kamat, A vision-based marker-less pose estimation system for articulated construction robots, Automation in Construction, 104 (2019) 80-94

    2019

  52. [60]

    Liang, S.-C

    C.-J. Liang, S.-C. Kang, M.-H. Lee, RAS: a robotic assembly system for steel structure erection and assembly, International Journal of Intelligent Robotics and Applications, 1 (2017) 459-476

    2017

  53. [61]

    Liang, V .R

    C.-J. Liang, V .R. Kamat, C.C. Menassa, W. McGee, Trajectory-based skill learning for overhead construction robots using generalized cylinders with orientation, Journal of Computing in Civil Engineering, 36 (2022) 04021036

    2022

  54. [62]

    Liang, V

    C.-J. Liang, V . Kamat, C. Menassa, Teaching robots to perform construction tasks via learning from demonstration, Proceedings of the 36th International Symposium on Automation and Robotics in Construction, ISARC, 2019, pp. 1305-1311

    2019

  55. [63]

    Asadi, V .R

    K. Asadi, V .R. Haritsa, K. Han, J.-P. Ore, Automated object manipulation using vision-based mobile robotic system for construction applications, Journal of Computing in Civil Engineering, 35 (2021) 04020058

    2021

  56. [64]

    Argall, B

    B. Argall, B. Browning, M. Veloso, Learning by demonstration with critique from a human teacher, Proceedings of the ACM/IEEE international conference on Human-robot interaction, 2007, pp. 57-64. 36

    2007

  57. [65]

    Karnan, G

    H. Karnan, G. Warnell, X. Xiao, P. Stone, V oila: Visual-observation-only imitation learning for autonomous navigation, 2022 International Conference on Robotics and Automation (ICRA), IEEE, 2022, pp. 2497-2503

    2022

  58. [66]

    H. Kim, Y . Ohmura, A. Nagakubo, Y . Kuniyoshi, Training robots without robots: deep imitation learning for master-to-robot policy transfer, IEEE Robotics and Automation Letters, 8 (2023) 2906-2913

    2023

  59. [67]

    Gandhi, S

    K. Gandhi, S. Karamcheti, M. Liao, D. Sadigh, Eliciting compatible demonstrations for multi-human imitation learning, Conference on Robot Learning, PMLR, 2023, pp. 1981-1991

    2023

  60. [68]

    O. Mees, L. Hermann, W. Burgard, What matters in language conditioned robotic imitation learning over unstructured data, IEEE Robotics and Automation Letters, 7 (2022) 11205-11212

    2022

  61. [69]

    B. Fang, S. Jia, D. Guo, M. Xu, S. Wen, F. Sun, Survey of imitation learning for robotic manipulation, International Journal of Intelligent Robotics and Applications, 3 (2019) 362-369

    2019

  62. [70]

    H. Yin, A. Varava, D. Kragic, Modeling, learning, perception, and control methods for deformable object manipulation, Science Robotics, 6 (2021) eabd8803

    2021

  63. [71]

    Q. Wu, X. Gong, K. Xu, D. Manocha, J. Dong, J. Wang, Towards target-driven visual navigation in indoor scenes via generative imitation learning, IEEE Robotics and Automation Letters, 6 (2020) 175-182

    2020

  64. [72]

    Doering, D.F

    M. Doering, D.F. Glas, H. Ishiguro, Modeling interaction structure for robot imitation learning of human social behavior, IEEE Transactions on Human-Machine Systems, 49 (2019) 219-231

    2019

  65. [73]

    R. Li, Z. Zou, Enhancing construction robot learning for collaborative and long-horizon tasks using generative adversarial imitation learning, Advanced Engineering Informatics, 58 (2023) 102140

    2023

  66. [74]

    Kappel, V

    M. Kappel, V . Golyanik, M. Elgharib, J.-O. Henningson, H.-P. Seidel, S. Castillo, C. Theobalt, M. Magnor, High-fidelity neural human motion transfer from monocular video, Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, 2021, pp. 1541-1550

    2021

  67. [75]

    S. Li, H. Yu, W. Ding, H. Liu, L. Ye, C. Xia, X. Wang, X.-P. Zhang, Visual–Tactile Fusion for Transparent Object Grasping in Complex Backgrounds, IEEE Transactions on Robotics, (2023)

    2023

  68. [76]

    Anzai, K

    T. Anzai, K. Takahashi, Deep gated multi-modal learning: In-hand object pose changes estimation using tactile and image data, 2020 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), IEEE, 2020, pp. 9361-9368

    2020

  69. [77]

    Boroushaki, J

    T. Boroushaki, J. Leng, I. Clester, A. Rodriguez, F. Adib, Robotic grasping of fully-occluded objects using rf perception, 2021 IEEE International Conference on Robotics and Automation (ICRA), IEEE, 2021, pp. 923-929

    2021

  70. [78]

    C. Yu, P. Wang, Dexterous manipulation for multi-fingered robotic hands with reinforcement learning: a review, Frontiers in Neurorobotics, 16 (2022) 861825

    2022

  71. [79]

    Yu, V .R

    H. Yu, V .R. Kamat, C.C. Menassa, W. McGee, Y . Guo, H. Lee, Mutual physical state-aware object handover in full-contact collaborative human-robot construction work, Automation in Construction, 150 (2023) 104829

    2023

  72. [80]

    Lambeta, P.-W

    M. Lambeta, P.-W. Chou, S. Tian, B. Yang, B. Maloon, V.R. Most, D. Stroud, R. Santos, A. Byagowi, G. Kammerer, Digit: A novel design for a low-cost compact high-resolution tactile sensor with application to in-hand manipulation, IEEE Robotics and Automation Letters, 5 (2020) 3838-3845

    2020

Pith tools

Reviewed August 10, 2026 · model on record in the stance chip above.