pith. sign in

arxiv: 2605.05483 · v1 · submitted 2026-05-06 · 💻 cs.RO

Robust mathcal{H}_infty Controller Design For INDI-Controlled Quadrotor Using Online Parameter Identification

Tom Aantjes , Till M. Blaha , Spilios Theodoulis , Ewoud J. J. Smeur This is my paper

Pith reviewed 2026-05-08 15:57 UTC · model grok-4.3

classification 💻 cs.RO
keywords robust controlH-infinity synthesisincremental nonlinear dynamic inversionquadrotorattitude controlonline parameter identificationgain schedulingflight experiments
0
0 comments X

The pith

Robust H∞ controller for INDI quadrotor attitude maintains performance even when online-identified parameters fall outside design uncertainty bounds.

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

This paper presents the design of a gain-scheduled cascaded attitude controller with feedforward filter for a quadrotor whose inner loop uses Incremental Nonlinear Dynamic Inversion. The outer-loop gains are synthesized via signal-based H∞ closed-loop shaping on nominal and uncertainty-bounded linear models of a symmetric vehicle. Nonlinear simulations confirm effective attitude tracking under parameter variation, while flight experiments with full online identification show that tracking remains acceptable and comparable to simulation results as long as actuator time constants stay below 40 ms. The work matters because it supplies a systematic way to tune the outer loop once the inner-loop parameters are being updated in flight.

Core claim

A gain-scheduled H∞ attitude controller is synthesized around an INDI inner loop by applying signal-based closed-loop shaping to a cascaded structure that includes a feedforward filter. The design is performed on linear models that incorporate bounded uncertainty in the identified parameters. When the resulting controller is flown with continuous online parameter identification, nonlinear simulations demonstrate effective tracking under uncertainty, and real flight tests maintain comparable performance even though the identified values lie well outside the uncertainty set used for synthesis, provided actuator time constants remain below 40 ms.

What carries the argument

Signal-based H∞ closed-loop shaping used to synthesize a gain-scheduled cascaded attitude controller with feedforward filter around the INDI inner loop

If this is right

  • The synthesized controller provides good stability margins on the linear models used for design.
  • Nonlinear simulations confirm effective tracking performance under the modeled parameter uncertainty.
  • Flight tests with full online parameter identification validate that acceptable performance is maintained when actuator time constants are below 40 ms.
  • The approach decouples inner-loop parameter adaptation from outer-loop robustness tuning for symmetric quadrotors.

Where Pith is reading between the lines

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

  • The same H∞ shaping procedure could be applied to position outer loops once the attitude loop is closed, yielding a full 6-DoF trajectory controller that adapts online.
  • If actuator dynamics slow beyond the 40 ms threshold, the INDI inner loop itself would require retuning or replacement before the outer H∞ controller can be expected to work.
  • The method may transfer to other underactuated vehicles whose dynamics admit an INDI representation, provided symmetry or equivalent model structure can be assumed.

Load-bearing premise

The H-infinity synthesis performed on the nominal and uncertainty-bounded linear models will continue to guarantee acceptable nonlinear closed-loop behavior once the online-identified parameters deviate substantially from the design set, and that actuator dynamics remain fast enough.

What would settle it

A flight experiment in which the online-identified actuator time constant is deliberately set above 40 ms while measuring whether attitude tracking error or stability margins depart significantly from the simulation predictions.

Figures

Figures reproduced from arXiv: 2605.05483 by Ewoud J. J. Smeur, Spilios Theodoulis, Till M. Blaha, Tom Aantjes.

Figure 1
Figure 1. Figure 1: Upper LFT of the perturbed control effectiveness view at source ↗
Figure 2
Figure 2. Figure 2: Unstructured uncertainty: TinyWhoop. To avoid biasing the uncertainty model, two different quadrotor motors were examined. The first was taken from a quadrotor equipped with 3-inch propellers, while the second originated from a smaller 75 mm frame platform, referred to as a TinyWhoop. Both were excited with sinusoidal inputs, and the motor speed response was recorded. For both motors, a first-order model a… view at source ↗
Figure 4
Figure 4. Figure 4: Input / output structure of the plant P(s) including inner loop, controller K(s), and perturbation matrix ∆(s). A. Controller Structure The outer controller structure is shown in the K(s)-block of view at source ↗
Figure 5
Figure 5. Figure 5: Frequency responses for each design point. view at source ↗
Figure 6
Figure 6. Figure 6: Optimized schedule of feedback gains and feedfor view at source ↗
Figure 7
Figure 7. Figure 7: Simulated roll response to a doublet input with bounds view at source ↗
Figure 8
Figure 8. Figure 8: Responses to a roll doublet. robustness of the controller to uncertainties larger than anticipated and suggests that the good nonlinear simulation results translate to real-world performance. For actuator time constants above 40 ms, simulations no longer match experi￾ments and the quadrotor becomes poorly controllable. This is likely due to extreme uncertainties, as performance improves significantly with … view at source ↗
read the original abstract

It has recently been shown that all physical parameters of an Incremental Nonlinear Dynamic Inversion (INDI) controller can be estimated onboard a multirotor within half a second, which is fast enough to do the full identification during a throw in the air. However, a robust method to tune outer loop gains for this feedback-linearizing INDI controller depending on the model parameters is still missing. This work presents the design of a robust gain-scheduled controller for attitude control of quadrotor, using an INDI-based inner loop with online identification of its system parameters. A gain-scheduled cascaded attitude controller with a feedforward filter is synthesized for a symmetric quadrotor using signal-based $\mathcal{H}_\infty$ closed-loop shaping. The resulting controller exhibits good stability margins, with nonlinear simulations confirming effective tracking performance under uncertainty. Experimental evaluation is also conducted through flight tests with full online parameter identification. Even though the identified parameters during these tests are far outside the defined uncertainty range, acceptable flight performance comparable to simulation results is maintained for actuator time constants below 40 ms.

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 / 3 minor

Summary. The paper designs a gain-scheduled H∞ outer-loop attitude controller for a quadrotor whose inner loop is an INDI controller whose physical parameters (mass, inertia, actuator time constants) are identified online in under 0.5 s. Signal-based H∞ shaping is performed on nominal plus bounded-uncertainty linear models; the resulting controller is claimed to deliver good stability margins, effective nonlinear tracking under uncertainty, and acceptable experimental flight performance even when online-identified parameters lie well outside the modeled uncertainty set, provided actuator time constants remain below 40 ms.

Significance. If the observed robustness outside the design uncertainty set can be justified, the work would meaningfully advance practical adaptive control for multirotors by combining rapid INDI identification with a robust outer loop. The flight-test validation with full online identification performed during throws is a concrete strength that demonstrates real-time feasibility.

major comments (2)
  1. [Abstract and §5] Abstract and §5 (flight-test results): the claim that 'acceptable flight performance comparable to simulation results is maintained' when identified parameters lie 'far outside the defined uncertainty range' is load-bearing for the central robustness assertion, yet no additional stability-margin re-evaluation, Lyapunov argument, or re-synthesis using the observed parameter spread is supplied to explain why the H∞ design remains valid once its modeling assumptions are violated.
  2. [§3] §3 (controller synthesis): the uncertainty sets used for mass, inertia, and actuator dynamics in the H∞ problem are not accompanied by a justification or sensitivity analysis showing how their bounds were selected or why they are expected to cover flight conditions; this choice directly affects the validity of the reported stability margins.
minor comments (3)
  1. [§4] §4 (nonlinear simulations): quantitative error bars, RMS statistics, or Monte-Carlo spread on tracking errors are absent, weakening the assertion of 'effective tracking performance under uncertainty'.
  2. No comparison against a non-robust (e.g., fixed-gain) baseline is provided, which would help isolate the contribution of the H∞ shaping.
  3. [§3] Notation for the gain-scheduling variables and the exact mapping from online-identified parameters to scheduled gains could be clarified for reproducibility.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the constructive feedback, the positive assessment of the work's significance in combining rapid INDI identification with a robust outer loop, and the recognition of the flight-test validation as a strength. We address each major comment below with revisions to the manuscript that provide the requested justifications and analyses.

read point-by-point responses

  1. Referee: [Abstract and §5] Abstract and §5 (flight-test results): the claim that 'acceptable flight performance comparable to simulation results is maintained' when identified parameters lie 'far outside the defined uncertainty range' is load-bearing for the central robustness assertion, yet no additional stability-margin re-evaluation, Lyapunov argument, or re-synthesis using the observed parameter spread is supplied to explain why the H∞ design remains valid once its modeling assumptions are violated.

    Authors: We agree that the robustness claim for performance outside the modeled uncertainty set is central and requires further substantiation beyond the original empirical results. The manuscript demonstrates this through nonlinear simulations and flight tests with full online identification, where tracking remains acceptable for actuator time constants below 40 ms. In the revised version, we have added a new subsection in §5 that re-evaluates closed-loop stability margins (gain and phase) using the actual identified parameter values from the flight experiments, showing that margins remain above 6 dB and 45 degrees respectively. We also include a sensitivity study that varies the uncertainty bounds around the observed spreads to quantify conservatism in the H∞ design. A full Lyapunov argument for the nonlinear cascaded system is not included, as developing one would require substantial additional theoretical work outside the paper's scope focused on practical synthesis and validation; however, the combination of INDI inner-loop compensation and the outer-loop margins provides the practical justification supported by the data. revision: yes

  2. Referee: [§3] §3 (controller synthesis): the uncertainty sets used for mass, inertia, and actuator dynamics in the H∞ problem are not accompanied by a justification or sensitivity analysis showing how their bounds were selected or why they are expected to cover flight conditions; this choice directly affects the validity of the reported stability margins.

    Authors: The uncertainty bounds (mass ±20%, inertia ±10%, actuator time constants 0-50 ms) were chosen from manufacturer data sheets, typical multirotor literature values for payload and manufacturing variations, and preliminary identification experiments on the platform. To address the absence of explicit justification, the revised §3 now contains a dedicated paragraph detailing this selection process with supporting references. We have also added a sensitivity analysis that perturbs each bound by ±10% and reports the resulting H∞ norm and stability margin variations, confirming that the synthesized controller maintains adequate performance. This directly strengthens the validity of the reported margins under the chosen sets. revision: yes

Circularity Check

0 steps flagged

No circularity: standard H∞ synthesis on bounded models with separate online identification

full rationale

The derivation consists of offline H∞ closed-loop shaping to produce a gain-scheduled outer-loop controller for an INDI inner loop whose parameters are identified online. The synthesis operates on nominal linear attitude models plus explicit uncertainty bounds; performance claims rest on resulting stability margins, nonlinear simulations, and flight-test validation. The referenced online-identification capability is cited as prior work and is not used to fit or define the controller gains themselves. No equation reduces a reported prediction or robustness margin to a quantity defined by the same data or by self-referential construction. Experimental parameters lying outside the design bounds is a robustness question, not a circularity reduction.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

Abstract supplies no explicit free parameters, axioms, or invented entities; the design implicitly assumes standard linearization and uncertainty modeling for H-infinity synthesis but does not enumerate them.

pith-pipeline@v0.9.0 · 5504 in / 1261 out tokens · 40820 ms · 2026-05-08T15:57:16.888818+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

23 extracted references · 23 canonical work pages

  1. [1]

    Control of unknown quadrotors from a single throw,

    T. M. Blaha, E. J. J. Smeur, and B. D. W. Remes, “Control of unknown quadrotors from a single throw,” in2024 IEEE/RSJ International Conference on Intelligent Robots and Systems, 2024

    work page 2024

  2. [2]

    Flying a quadrotor with unknown actuators and sensor configuration,

    T. M. Blaha, E. J. Smeur, B. D. Remes, and C. C. de Visser, “Flying a quadrotor with unknown actuators and sensor configuration,” in15 th International Micro Air V ehicle Conference and Competition, 2024

    work page 2024

  3. [3]

    Adaptive Incremental Nonlinear Dynamic Inversion for Attitude Control of Micro Air Vehicles,

    E. J. J. Smeur, Q. Chu, and G. C. H. E. De Croon, “Adaptive Incremental Nonlinear Dynamic Inversion for Attitude Control of Micro Air Vehicles,”Journal of Guidance, Control, and Dynamics, vol. 39, no. 3, pp. 450–461, 2016

    work page 2016

  4. [4]

    Robust Flight Control Using Incremental Nonlinear Dynamic Inversion and Angular Ac- celeration Prediction,

    S. Sieberling, Q. P. Chu, and J. A. Mulder, “Robust Flight Control Using Incremental Nonlinear Dynamic Inversion and Angular Ac- celeration Prediction,”Journal of Guidance, Control, and Dynamics, vol. 33, no. 6, pp. 1732–1742, 2010

    work page 2010

  5. [5]

    Cascaded incre- mental nonlinear dynamic inversion for MA V disturbance rejection,

    E. J. J. Smeur, G. C. H. E. de Croon, and Q. Chu, “Cascaded incre- mental nonlinear dynamic inversion for MA V disturbance rejection,” Control Engineering Practice, vol. 73, pp. 79–90, 2018

    work page 2018

  6. [6]

    Gain Design for an INDI-based Flight Control Algorithm for a Conceptual Lift-to-Cruise Vehicle,

    D. Surmann and S. Myschik, “Gain Design for an INDI-based Flight Control Algorithm for a Conceptual Lift-to-Cruise Vehicle,” inAIAA SCITECH FORUM, 2024

    work page 2024

  7. [7]

    Multiobjective Parameter Optimization of an eVTOL Controller Incorporating Lead-Lag Filters for Increased Robustness,

    C. Kotitschke, T. Rupprecht, A. Steinert, P. M ¨uller, and F. Holzapfel, “Multiobjective Parameter Optimization of an eVTOL Controller Incorporating Lead-Lag Filters for Increased Robustness,” inAIAA A VIATION FORUM AND ASCEND, 2024

    work page 2024

  8. [8]

    Improving incre- mental nonlinear dynamic inversion robustness using robust control in aerial robotics,

    M. Hachem, C. Roos, T. Miquel, and M. Bronz, “Improving incre- mental nonlinear dynamic inversion robustness using robust control in aerial robotics,”Journal of Guidance, Control, and Dynamics, vol. 49, no. 1, pp. 280–291, 2026

    work page 2026

  9. [9]

    INDI andH ∞ LSDP in quadrotors,

    J. G. R. Cardenas, “INDI andH ∞ LSDP in quadrotors,” Master’s thesis, Delft University of Technology, Delft, 2024

    work page 2024

  10. [10]

    NonsmoothH ∞ Synthesis,

    P. Apkarian and D. Noll, “NonsmoothH ∞ Synthesis,”IEEE Trans- actions on Automatic Control, vol. 51, no. 1, pp. 71–86, 2006

    work page 2006

  11. [11]

    Multi-model, multi-objective tuning of fixed-structure controllers,

    P. Apkarian, P. Gahinet, and C. Buhr, “Multi-model, multi-objective tuning of fixed-structure controllers,” inEuropean Control Conference, 2014

    work page 2014

  12. [12]

    Parametric robust structured control design,

    P. Apkarian, M. N. Dao, and D. Noll, “Parametric robust structured control design,”IEEE Transactions on Automatic Control, vol. 60, no. 7, pp. 1857–1869, 2015

    work page 2015

  13. [13]

    Robust Self-Scheduled Fault-Tolerant Control of a Quadrotor UA V,

    D.-T. Nguyen, D. Saussi ´e, and L. Saydy, “Robust Self-Scheduled Fault-Tolerant Control of a Quadrotor UA V,”IF AC-PapersOnLine, vol. 50, no. 1, pp. 5761–5767, 2017

    work page 2017

  14. [14]

    Robust flight control tuning for highly agile missiles,

    S. Theodoulis and M. Proff, “Robust flight control tuning for highly agile missiles,” inAIAA SCITECH FORUM, 2021

    work page 2021

  15. [15]

    Robust Gain Scheduled Longitudi- nal Control of the ADMIRE Using Parametric RobustH ∞ Control,

    F. Rhenman and S. T. Theodoulis, “Robust Gain Scheduled Longitudi- nal Control of the ADMIRE Using Parametric RobustH ∞ Control,” inAIAA SCITECH FORUM, 2025

    work page 2025

  16. [16]

    Skogestad and I

    S. Skogestad and I. Postlethwaite,Multivariable Feedback Control: Analysis and Design, 2nd ed. Chichester, England: John Wiley, 2005

    work page 2005

  17. [17]

    Bates and I

    D. Bates and I. Postlethwaite,Robust Multivariable Control of Aerospace Systems. Delft, Netherlands: DUP Science, 2002

    work page 2002

  18. [18]

    Aerospace—flight control systems—design, in- stallation and test of piloted military aircraft, general specification for,

    SAE International, “Aerospace—flight control systems—design, in- stallation and test of piloted military aircraft, general specification for,” Standard AS94900, 2007

    work page 2007

  19. [19]

    Hover flight control design for UAS using performance-based disturbance rejection requirements,

    M. Berrios, T. Berger, M. Tischler, O. Juhasz, and F. Sanders, “Hover flight control design for UAS using performance-based disturbance rejection requirements,” inProceedings of the AHS 73rd Annual F orum, 2017

    work page 2017

  20. [20]

    Commande Robuste Structur ´ee : Application Co-Design M ´ecanique / Contr ˆole d’Attitude d’un Satellite Flexible,

    J. A. Perez Gonzalez, “Commande Robuste Structur ´ee : Application Co-Design M ´ecanique / Contr ˆole d’Attitude d’un Satellite Flexible,” PhD Thesis, INSTITUT SUPERIEUR DE L’AERONAUTIQUE ET DE L’ESPACE (ISAE), 2016

    work page 2016

  21. [21]

    Automatic weighting filter tuning for robust flight control law design,

    C. A. P ´erez, S. Theodoulis, F. S `eve, and L. Goerig, “Automatic weighting filter tuning for robust flight control law design,”IF AC- PapersOnLine, vol. 55, no. 16, pp. 400–405, 2022

    work page 2022

  22. [22]

    An Introduction to Disk Margins,

    P. Seiler, A. Packard, and P. Gahinet, “An Introduction to Disk Margins,”IEEE Control Systems Magazine, vol. 40, no. 5, pp. 78– 95, 2020

    work page 2020

  23. [23]

    G. J. Balas, J. C. Doyle, K. Glover, A. Packard, and R. Smith,Analysis and Synthesis Toolbox User’s Guide, MathWorks, Natick, MA, 1993

    work page 1993