pith. sign in

arxiv: 2604.18905 · v1 · submitted 2026-04-20 · 💻 cs.RO

Task-Adaptive Admittance Control for Human-Quadrotor Cooperative Load Transportation with Dynamic Cable-Length Regulation

Shuai Li , Ton T. H. Duong , Damiano Zanotto This is my paper

Pith reviewed 2026-05-10 03:42 UTC · model grok-4.3

classification 💻 cs.RO
keywords admittance controlcooperative load transportationhuman-quadrotor interactionvariable cable lengthphysical human-robot interactionaerial manipulationwinch regulation
0
0 comments X

The pith

A quadrotor with an active winch uses admittance control that incorporates coupled dynamics to adapt cable length in real time to human contact forces during load transport.

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

This paper presents an admittance controller for a quadrotor that works with a human to move loads while an onboard winch changes cable length on the fly. The controller treats the drone, cable, and human-applied forces as a single interacting system so the cable can shorten or lengthen to match changing contact conditions. Experiments test the full sequence of picking up a load on the ground, carrying it, and setting it down, comparing the new method against standard fixed-length control under both soft and stiff contact. The results show faster response to human pushes and smoother overall motion. Readers would care because the approach points toward safer, more natural physical teamwork between people and flying robots for tasks such as moving furniture or equipment in tight spaces.

Core claim

The proposed admittance controller accounts for the coupled dynamics of the quadrotor, variable-length cable, and human-applied forces, allowing the system to dynamically adapt cable length and respond to contact during the entire cooperative load transportation process, which produces better responsiveness and smoother trajectories than conventional fixed-cable admittance control.

What carries the argument

The task-adaptive admittance controller with dynamic cable-length regulation, which continuously adjusts winch commands by modeling the combined quadrotor-cable-human force dynamics to match task phases such as loading, transport, and unloading.

Load-bearing premise

The combined dynamics of the quadrotor, the changing cable length, and the forces applied by the human can be modeled accurately enough and sensed fast enough for real-time control without large unmodeled effects or delays.

What would settle it

A side-by-side experiment in which the proposed controller produces no measurable improvement in response time to sudden human force changes or in trajectory smoothness compared with a standard fixed-length admittance controller under high-stiffness contact would show the adaptation claim does not hold.

Figures

Figures reproduced from arXiv: 2604.18905 by Damiano Zanotto, Shuai Li, Ton T. H. Duong.

Figure 1
Figure 1. Figure 1: (a) Proposed human-quadrotor CLT system with active cable length control. The obstacle (truncated in height, for better visualization), is posi￾tioned along the load’s path to the destination, prompting the human operator to take action to alter the system’s motion to avoid the obstacle. This is safely achieved using an admittance controller. (b) Model of the CLT system, showing the inertial frame {I}, wit… view at source ↗
Figure 2
Figure 2. Figure 2: (a) In the coupled virtual impedance model (CVIM), the admittance error eac is proportional to the difference between the virtual load direction e v l and its desired direction e d l (green arrow), which also represents the required quadrotor translational motion (blue dashed arrow). (b) In the simplified virtual impedance model (SVIM), the compliant motion during interaction is separated into two componen… view at source ↗
Figure 3
Figure 3. Figure 3: Proposed control scheme for the human–quadrotor CLT tasks. [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: (a) Loading/Unloading in Place – this task requires the operator to attach the load to the cable hook and guide the CLT system to a location as far away as possible from the starting position, within a designated time window; (b) Transporting – this task encompasses the collaborative transportation of an object to a target destination, while the operator’s action avoids collision between the CLT system and… view at source ↗
Figure 5
Figure 5. Figure 5: Loading/Unloading in Place task: (a) Traversed distance d, (b) inclination angle β¯, (c) smoothness of the drone trajectory J¯q, and (d) cable tensile force f¯ T , averaged across ten trials, for each combination of control method, cable configuration, and stiffness level. Transporting task: (e) Normalized travel distance dn, (f) β¯, (g) J¯q, and (h) f¯ T , averaged across ten trials for each combination o… view at source ↗
Figure 6
Figure 6. Figure 6: Top views of the virtual references (dashed orange lines) and operator￾guided paths (solid blue lines) followed by the quadrotor’s center of mass during representative transport tasks. Green solid lines and red arrows indicate the actual path traversed by the load (represented by the cable’s distal point) and the contact force Fc exerted by the operator to avoid collisions with the obstacle (light blue rec… view at source ↗
Figure 7
Figure 7. Figure 7: (a) Norm of the admittance error eac for CVIM, under each combination of cable configuration and stiffness level; (b) Corresponding cable-length adjustment δL¯ for the VCL configurations [PITH_FULL_IMAGE:figures/full_fig_p007_7.png] view at source ↗
Figure 8
Figure 8. Figure 8: Contact force Fc obtained with CVIM for each cable configuration and stiffness level. The components of Fc are reported along the axes of the inertial frame {I} defined in [PITH_FULL_IMAGE:figures/full_fig_p007_8.png] view at source ↗
read the original abstract

The collaboration between humans and robots is critical in many robotic applications, especially in those requiring physical human-robot interaction (pHRI). Previous research in pHRI has largely focused on robotic manipulators, employing impedance or admittance control to maintain operational safety. Conversely, research in human-quadrotor cooperative load transportation (CLT) is still in its infancy. This letter introduces a novel admittance controller designed for safe and effective human-quadrotor CLT using a quadrotor equipped with an actively-controlled winch. The proposed method accounts for the system's coupled dynamics, allowing the quadrotor and its cable to dynamically adapt to contact forces during CLT tasks, thereby enhancing responsiveness. We experimentally validated the task-adaptive capability of the controller across the entire CLT process, including in-place loading/unloading and load transporting tasks. To this end, we compared the system performances against a conventional approach, using both variable and fixed cable lengths under low- and high-stiffness conditions. Results demonstrate that the proposed method outperforms the conventional approach in terms of system responsiveness and motion smoothness, leading to improved CLT capabilities.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

2 major / 0 minor

Summary. The paper proposes a task-adaptive admittance controller for human-quadrotor cooperative load transportation (CLT) using a quadrotor with an actively controlled winch. The controller incorporates the coupled dynamics of the quadrotor, variable-length cable, and human-applied forces to adapt to contact forces during in-place loading/unloading and transport tasks. Experimental comparisons against conventional fixed- and variable-length cable approaches under low- and high-stiffness conditions are reported to demonstrate improved system responsiveness and motion smoothness.

Significance. If the central claims hold with supporting data, the work would represent a meaningful extension of admittance control principles to aerial pHRI, addressing an emerging application area where prior focus has been on manipulators. The task-adaptive formulation and full-process experimental validation (loading through transport) are potentially useful strengths for practical deployment in logistics or assistance scenarios. However, the absence of quantitative metrics, error statistics, or explicit controller equations in the provided manuscript text limits the ability to assess the magnitude of improvement or generalizability.

major comments (2)
  1. [Abstract] Abstract: The central claim that the proposed method 'outperforms the conventional approach in terms of system responsiveness and motion smoothness' is stated without any quantitative metrics, error bars, statistical tests, or specific values (e.g., settling time, RMS position error, or force variance). This is load-bearing for the experimental validation of the task-adaptive capability.
  2. [Abstract] Abstract: The statement that the method 'accounts for the system's coupled dynamics' is central to the novelty, yet no details are provided on the dynamic model, real-time force sensing, cable tension formulation, or the explicit admittance law (including any adaptation mechanism for cable length). Without these, it is not possible to evaluate whether unmodeled disturbances or delays are adequately handled, as required by the weakest assumption in the work.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the constructive comments on our manuscript. We agree that the abstract would benefit from additional quantitative details and clearer pointers to the technical content. We will revise the abstract accordingly while preserving the letter format. Point-by-point responses follow.

read point-by-point responses

  1. Referee: [Abstract] Abstract: The central claim that the proposed method 'outperforms the conventional approach in terms of system responsiveness and motion smoothness' is stated without any quantitative metrics, error bars, statistical tests, or specific values (e.g., settling time, RMS position error, or force variance). This is load-bearing for the experimental validation of the task-adaptive capability.

    Authors: The manuscript body reports quantitative experimental comparisons (position RMS errors, force variance, settling times, and smoothness metrics) against fixed- and variable-length baselines under both low- and high-stiffness conditions. To address the concern, we will augment the abstract with specific numerical improvements drawn from those results (e.g., percentage reductions in RMS position error and settling time) while retaining the high-level summary style appropriate for a letter. revision: yes

  2. Referee: [Abstract] Abstract: The statement that the method 'accounts for the system's coupled dynamics' is central to the novelty, yet no details are provided on the dynamic model, real-time force sensing, cable tension formulation, or the explicit admittance law (including any adaptation mechanism for cable length). Without these, it is not possible to evaluate whether unmodeled disturbances or delays are adequately handled, as required by the weakest assumption in the work.

    Authors: The full manuscript derives the coupled quadrotor-cable-human dynamics in Section II and presents the task-adaptive admittance law with explicit cable-length regulation in Section III, including onboard load-cell force sensing and tension modeling. We will revise the abstract to include a concise reference to the coupled dynamic model and real-time force feedback. The explicit equations remain in the main text for full evaluation of disturbance handling. revision: partial

Circularity Check

0 steps flagged

No significant circularity in derivation chain

full rationale

The paper extends established admittance control principles to a human-quadrotor CLT system with active winch and variable cable length, claiming improved responsiveness by accounting for coupled dynamics. No derivation steps, equations, or self-citations are shown that reduce predictions to fitted inputs by construction, self-define variables in terms of outputs, or rely on load-bearing uniqueness theorems from the same authors. Experimental comparisons to fixed/variable-length baselines provide independent validation content, keeping the central claim self-contained against external benchmarks.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 0 invented entities

Review based solely on abstract; controller design implicitly rests on standard robotics assumptions about accurate real-time force sensing and model-based adaptation of coupled dynamics, with no explicit free parameters or invented entities stated.

axioms (1)
  • domain assumption The quadrotor-cable-human interaction can be modeled as coupled dynamics that admit stable admittance control adaptation across task phases.
    Invoked to justify dynamic cable regulation and responsiveness gains during loading, transport, and unloading.

pith-pipeline@v0.9.0 · 5507 in / 1291 out tokens · 31098 ms · 2026-05-10T03:42:53.376793+00:00 · methodology

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

28 extracted references · 28 canonical work pages

  1. [1]

    Wing expands service to melbourne with doordash while advancing australian regulatory approvals,

    M. Nagle, “Wing expands service to melbourne with doordash while advancing australian regulatory approvals,”Wing Blog, 2024. [Online]. Available: https://wing.com/news/wing-melbourne-doordash

    work page 2024

  2. [2]

    Last-mile drone delivery service market to surpass USD 2,582.8 million by 2033,

    Y . Shinde, “Last-mile drone delivery service market to surpass USD 2,582.8 million by 2033,”market.us, 2025. [Online]. Available: https://market.us/report/last-mile-drone-delivery-service-market/

    work page 2033

  3. [3]

    Amazon Drone Crashes Hit Jeff Bezos’ Delivery Dreams,

    S. Soper and M. Day, “Amazon Drone Crashes Hit Jeff Bezos’ Delivery Dreams,”Bloomberg, 2022. [On- line]. Available: https://www.bloomberg.com/news/features/2022-04-10/ amazon-drone-crashes-delays-put-bezos-s-delivery-dream-at-risk

    work page 2022

  4. [4]

    NASA and the FAA are working to revolutionize air traffic control for the drone era,

    C. F. Arnold, “NASA and the FAA are working to revolutionize air traffic control for the drone era,”Bloomberg, 2022. [Online]. Available: https://www.bloomberg.com/features/2022-drone-traffic-control-nasa/

    work page 2022

  5. [5]

    Toward physical human-robot interaction control with aerial manipulators: Compliance, redundancy resolution, and input limits,

    A. Afifi, M. van Holland, and A. Franchi, “Toward physical human-robot interaction control with aerial manipulators: Compliance, redundancy resolution, and input limits,” inIEEE ICRA, 2022

    work page 2022

  6. [6]

    Admittance force- based UA V-wall stabilization and press exertion for documentation and inspection of historical buildings,

    D. Smrcka, T. Baca, T. Nascimento, and M. Saska, “Admittance force- based UA V-wall stabilization and press exertion for documentation and inspection of historical buildings,” inICUAS, 2021

    work page 2021

  7. [7]

    Image-based visual- impedance control of a dual-arm aerial manipulator,

    V . Lippiello, G. A. Fontanelli, and F. Ruggiero, “Image-based visual- impedance control of a dual-arm aerial manipulator,”IEEE RA-L, vol. 3, no. 3, pp. 1856–1863, 2018

    work page 2018

  8. [8]

    An aerial parallel manipulator with shared compliance,

    B. Stephens, L. Orr, B. B. Kocer, H.-N. Nguyen, and M. Kovac, “An aerial parallel manipulator with shared compliance,”IEEE RA-L, vol. 7, no. 4, pp. 11 902–11 909, 2022

    work page 2022

  9. [9]

    Tracking control of fully-constrained cable-driven parallel robots using adaptive dynamic programming,

    S. Li and D. Zanotto, “Tracking control of fully-constrained cable-driven parallel robots using adaptive dynamic programming,” inIEEE IROS, 2019, pp. 6781–6787

    work page 2019

  10. [10]

    Multi-UA V cooperative transportation using dynamic control allocation and a reinforcement learning compensator,

    ——, “Multi-UA V cooperative transportation using dynamic control allocation and a reinforcement learning compensator,” inASME IDETC, vol. 85468, 2021, p. V009T09A034

    work page 2021

  11. [11]

    In-flight cable length control for improved quadrotor-based suspended load transportation,

    S. Li, T. T. H. Duong, and D. Zanotto, “In-flight cable length control for improved quadrotor-based suspended load transportation,”IEEE RA-L, vol. 9, no. 1, pp. 667–674, 2024

    work page 2024

  12. [12]

    Physical human-robot interac- tion with a tethered aerial vehicle: Application to a force-based human guiding problem,

    M. Tognon, R. Alami, and B. Siciliano, “Physical human-robot interac- tion with a tethered aerial vehicle: Application to a force-based human guiding problem,”IEEE TRO, vol. 37, no. 3, pp. 723–734, 2021

    work page 2021

  13. [13]

    Human-state-aware controller for a tethered aerial robot guiding a human by physical interaction,

    M. Allenspach, Y . Vyas, M. Rubio, R. Siegwart, and M. Tognon, “Human-state-aware controller for a tethered aerial robot guiding a human by physical interaction,”IEEE RA-L, vol. 7, no. 2, pp. 2827– 2834, 2022

    work page 2022

  14. [14]

    Visual impedance based human-robot co-transportation with a tethered aerial vehicle,

    M. Xu, A. Hu, and H. Wang, “Visual impedance based human-robot co-transportation with a tethered aerial vehicle,”IEEE TII, 2023

    work page 2023

  15. [15]

    Aerial physical human robot interaction for payload transportation,

    P. Prajapati and V . Vashista, “Aerial physical human robot interaction for payload transportation,”IEEE RA-L, 2023

    work page 2023

  16. [16]

    Cooperative multilift slung load transportation using haptic admittance control guidance,

    M. Romano, A. Ye, J. Pye, and E. Atkins, “Cooperative multilift slung load transportation using haptic admittance control guidance,”JGCD, vol. 45, no. 10, pp. 1899–1912, 2022

    work page 1912

  17. [17]

    Human-aware physical human-robot col- laborative transportation and manipulation with multiple aerial robots,

    G. Li, X. Liu, and G. Loianno, “Human-aware physical human-robot col- laborative transportation and manipulation with multiple aerial robots,” IEEE TRO, 2024

    work page 2024

  18. [18]

    Geometric control and differential flatness of a quadrotor uav with load suspended from a pulley,

    J. Zeng, P. Kotaru, and K. Sreenath, “Geometric control and differential flatness of a quadrotor uav with load suspended from a pulley,” inACC, 2019, pp. 2420–2427

    work page 2019

  19. [19]

    Admittance control for physical human–robot interaction,

    A. Q. Keemink, H. Van der Kooij, and A. H. Stienen, “Admittance control for physical human–robot interaction,”IJR, vol. 37, no. 11, pp. 1421–1444, 2018

    work page 2018

  20. [20]

    Geometric control and differential flatness of a quadrotor uav with a cable-suspended load,

    K. Sreenath, T. Lee, and V . Kumar, “Geometric control and differential flatness of a quadrotor uav with a cable-suspended load,” inIEEE CDC, 2013, pp. 2269–2274

    work page 2013

  21. [21]

    A review of real- time implementable cooperative aerial manipulation systems,

    S. C. Barakou, C. S. Tzafestas, and K. P. Valavanis, “A review of real- time implementable cooperative aerial manipulation systems,”Drones, vol. 8, no. 5, p. 196, 2024

    work page 2024

  22. [22]

    A coordinated framework of aerial manipulator for safe and compliant physical inter- action,

    Q. Liu, S. Lyu, K. Guo, J. Wang, X. Yu, and L. Guo, “A coordinated framework of aerial manipulator for safe and compliant physical inter- action,”CEP, vol. 146, p. 105898, 2024

    work page 2024

  23. [23]

    Robust collaborative object transportation using multiple MA Vs,

    A. Tagliabue, M. Kamel, R. Siegwart, and J. Nieto, “Robust collaborative object transportation using multiple MA Vs,”IJR, vol. 38, no. 9, pp. 1020–1044, 2019

    work page 2019

  24. [24]

    HPA-MPC: Hybrid Perception- Aware Nonlinear Model Predictive Control for Quadrotors with Sus- pended Loads,

    M. Sarvaiya, G. Li, and G. Loianno, “HPA-MPC: Hybrid Perception- Aware Nonlinear Model Predictive Control for Quadrotors with Sus- pended Loads,”IEEE RA-L, 2024

    work page 2024

  25. [25]

    Enhancing safety in collaborative cable-driven parallel robots: Contact distinction and management for carrying tasks,

    H. Gao, C. Chevallereau, and S. Caro, “Enhancing safety in collaborative cable-driven parallel robots: Contact distinction and management for carrying tasks,”IEEE TASE, 2025. LIet al.: TASK-ADAPTIVE ADMITTANCE CONTROL FOR HUMAN-QUADROTOR COOPERATIVE LOAD TRANSPORTATION 9 APPENDIXA STABILITYANALYSIS A. Stability of the Admittance Controller Theorem 1.Consi...

    work page 2025

  26. [26]

    Letz= [ζ T , ˙ζ T ]T

    The admittance system is input-to-state stable (ISS) with respect toτ c(t). Letz= [ζ T , ˙ζ T ]T . There exist positive constantsc 1,c 2,λ 1, andΓsuch that ∥z(t)∥2≤ c2 c1 e−λt∥z(0)∥2+ Γ λc1 sup 0≤s≤t ∥τ c(s)∥2.(18)

    work page

  27. [27]

    Ifτ c(t)≡0fort≥T c, there exists a positive constantλ 2, such that ∥z(t)∥2≤ c2 c1 e−λ2(t−Tc)∥z(Tc)∥2, t≥T c.(19)

    The admittance system is exponentially stable during interaction. Ifτ c(t)≡0fort≥T c, there exists a positive constantλ 2, such that ∥z(t)∥2≤ c2 c1 e−λ2(t−Tc)∥z(Tc)∥2, t≥T c.(19)

    work page

  28. [28]

    Proof.A Lyapunov candidate function is chosen as Vac = 1 2 ζT Kζ+ 1 2 ˙ζ T M(t)ζ+ϵζ T K ˙ζ,(20) withϵ >0

    The command-shaping outputs are bounded for allt and exponentially converge to zero once the interaction vanishes. Proof.A Lyapunov candidate function is chosen as Vac = 1 2 ζT Kζ+ 1 2 ˙ζ T M(t)ζ+ϵζ T K ˙ζ,(20) withϵ >0. SinceM(t)andKare diagonal,V ac can be expressed in block-diagonal form as Vac = 1 2 [δβ, δ ˙β]Qβ[δβ, δ ˙β]T + [δL, δ ˙L]QL[δL, δ ˙L]T , ...

    work page 2026