Pith. sign in

REVIEW 3 major objections 5 minor 1 cited by

A Koopman-based adaptive digital twin can steer sheet metal forming to a target shape by learning a linear lifted-space model and updating its control matrix online.

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

2026-08-03 22:20 UTC pith:O5BSJYUQ

load-bearing objection A genuinely novel Koopman-MPC pipeline for sheet metal forming with real closed-loop experiments, but the headline claim overstates what the large-deformation results show. the 3 major comments →

arxiv 2511.10852 v1 pith:O5BSJYUQ submitted 2025-11-13 eess.SY cs.AIcs.SY

Adaptive Digital Twin of Sheet Metal Forming via Proper Orthogonal Decomposition-Based Koopman Operator with Model Predictive Control

Yi-Ping Chen , Derick Suarez , Ying-Kuan Tsai , Vispi Karkaria , Guanzhong Hu , Zihan Chen , Ping Guo , Jian Cao
show 1 more author
Wei Chen
This is my paper
classification eess.SY cs.AIcs.SY
keywords Digital TwinSheet Metal FormingEnglish WheelKoopman OperatorModel Predictive ControlProper Orthogonal DecompositionOnline Model UpdatingRecursive Least Squares
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.

The paper tries to establish that an adaptive digital twin can autonomously plan and correct toolpaths for a sheet-metal forming process (the English wheel) where both state and input are high-dimensional spatial fields. It does so by compressing shapes and toolpaths, fitting a Koopman operator that makes the nonlinear evolution linear in a lifted space, and then solving model predictive control as a convex program. Crucially, the control matrix is updated online with recursive least squares to counter material hardening and drift. Experiments show the adaptive system reduces maximum final deviation from 12.64 mm to 2.79 mm for a large-deformation target and from 7.34 mm to 1.09 mm for a moderate one. If correct, this offers a template for real-time autonomous control of deformation-based manufacturing processes.

Core claim

The paper shows experimentally that a robotic English wheel can be controlled autonomously by a digital twin built from a POD-reduced deep Koopman operator. Deformation fields are compressed to four PCA coefficients, toolpaths to five Chebyshev coefficients, and a neural network learns observables that make the evolution linear: z(k+1) = A z(k) + B u(k). Model predictive control then becomes a convex quadratic program solved in about 0.8 seconds. Because the material hardens and the static model over-predicts toolpath effect, only the B matrix (the control-effectiveness map) is re-estimated online via recursive least squares when prediction error exceeds 3 mm or deformation stagnates. With u

What carries the argument

The central object is the finite-dimensional Koopman operator with control, A and B acting in a lifted space. A is the state-evolution matrix and B maps Chebyshev toolpath coefficients to the lifted state; together they convert the nonlinear forming dynamics into a linear system usable for convex MPC. POD provides the spatial basis that keeps the state small, and residual-connected encoder/decoder networks learn the lifting function psi. The online piece is recursive least squares acting only on B, with a Kalman gain P and forgetting factor lambda = 0.9; this compensates for strain hardening without retraining the network.

Load-bearing premise

The next deformed shape depends only on the current measured shape and the chosen toolpath; the material's hardening and stress state are not part of the state, so if two sheets with identical measured geometry but different deformation histories respond differently to the same toolpath, the model is misspecified.

What would settle it

Take two sheets with the same initial measured midline shape, plastically pre-strain one of them, then run the same nominal toolpath on both and compare the post-cycle deformation fields; if they differ substantially, the first-order Markov state is insufficient and the B-only online update cannot fully absorb the gap.

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

If this is right

  • Closed-loop autonomous forming becomes practical: toolpaths are computed in under a second and the sheet reaches its target shape without human adjustment.
  • The pipeline transfers to other deformation processes, such as incremental forming, forging, or additive manufacturing, wherever state and input are spatial fields and data are scarce.
  • Online updating of B is sufficient to offset material drift in these experiments, suggesting that global retraining can be avoided for non-stationary process conditions.
  • The Koopman structure yields interpretable control gains: the observed reduction in the T1-to-phi0 coefficient shows which physical channel becomes less effective as the metal hardens.

Where Pith is reading between the lines

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

  • If material history, for example accumulated effective strain, were added as an explicit state, the model would likely need smaller B corrections and could generalize to new sequences without additional updating.
  • The fixed-A / updated-B strategy should be stress-tested on a different material or sheet thickness; a failure there would indicate the need to update A or the observables as well.
  • A natural next experiment is active data collection: use the Fisher information of A and B to plan toolpaths that reduce model uncertainty, which the paper itself names as future work.
  • Extending the deformation representation from a one-dimensional midline to a two-dimensional surface basis would let the same framework control full 3D panel shapes with the same MPC formulation.

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

3 major / 5 minor

Summary. The paper proposes an adaptive digital twin framework for robotic English wheel sheet metal forming. A POD-based deep Koopman operator learns a reduced-order linear lifted-state model from experimental deformation fields and toolpaths; the resulting linear model is embedded in a constrained MPC that designs six-cycle toolpath sequences. An RLS module updates the control matrix B online to compensate for unmodeled material-state effects. The framework is validated on 80 experimental forming runs for surrogate identification and on two closed-loop control cases (large and moderate target deformations), each compared with and without online model updating. The reported results show large improvements: maximum final deviation falls from 12.64 mm to 2.79 mm in the large-deformation case and from 7.34 mm to 1.09 mm in the moderate-deformation case.

Significance. If the claims are supported, the work is a meaningful experimental demonstration of closed-loop, model-based control for a nonlinear, high-dimensional manufacturing process using a learned reduced-order Koopman model with online adaptation. The paper ships physical experiments rather than simulation-only evidence, and the comparison with/without updating is a useful ablation. The POD-plus-Chebyshev decomposition is a sensible way to keep both states and inputs low-dimensional while preserving spatial structure and toolpath smoothness. However, the central claim that the adaptive DT can 'achieve the given target shape' is not supported by the paper's own termination criterion in the large-deformation experiment, and the closed-loop evidence consists of single runs without repeated trials or error bars despite a quoted process repeatability of ±1.5 mm. These gaps need to be addressed before the headline claim can be accepted as stated.

major comments (3)
  1. [Abstract; §5.1; §5.3, Fig. 9] The abstract states that the adaptive DT is 'capable of controlling the forming process to achieve the given target shape.' In the large-deformation experiment with online updating, the final maximum deviation is 2.79 mm at tracker location 7 (Figure 9), while §5.1 defines the termination/success criterion as max deviation below 1.5 mm. Thus the large-deformation case does not achieve the target under the paper's own tolerance; it substantially reduces the deviation from 12.64 mm. The moderate-deformation case does reach 1.09 mm and terminates in five cycles, so the claim is supported only for that case. This evidence-to-claim gap is load-bearing because target achievement is the headline contribution. Please qualify the claim (e.g., 'substantially reduce deviation' or 'approach the target within a specified tolerance') and explicitly state which cases meet the 1.5 mm criterion.
  2. [§5.3, Figs. 9–11] All closed-loop control results are single experimental runs. The paper itself quotes a process repeatability of ±1.5 mm at the free end (§2). With one run per condition, there is no way to assess run-to-run variability, sensitivity to initial conditions, or whether the reported improvements are typical. The improvement magnitudes (12.64→2.79 mm and 7.34→1.09 mm) exceed the repeatability, which is encouraging, but a control-validation claim of this strength should be supported by repeated trials (at least three per condition) or a statistical treatment. This is not fatal to the methodology, but it is necessary for the stated generalizability claim in the abstract and Closure.
  3. [§5.2, Eqs. (21)–(24)] The RLS update attributes the entire one-step prediction residual e_k = z*_{k+1} - A z_k to the control matrix B, while A, the encoder, and the decoder remain fixed. The paper acknowledges that material state is not included as an input (§5.2), so the residual will also contain unmodeled state-dependent dynamics. The blame assignment to B is therefore a strong modeling assumption. The experimental improvement suggests the heuristic is useful, but the paper does not test whether updating A or the full model would perform differently, nor does it discuss conditions under which fixing A could distort the adapted model. A comparative study or at least a careful discussion of this identifiability issue would strengthen the 'adaptive digital twin' claim.
minor comments (5)
  1. [Throughout] Typographical and language issues: 'dyanmics' (§2), 'elaborated in in Section' (§2), 'Minstry' (Acknowledgement), 'feasability' (Ref. [15]), 'Adaptive ratio' for λ in Eq. (22)–(24), and inconsistent use of 'data' as singular/plural.
  2. [Fig. 6 caption] The caption for Figure 6 lists subplots '(a) Comparison ... (d) Reconstruction error'; the second subplot should be labeled '(b)' to match the text reference 'Figure 6(b)'.
  3. [§4.4, Fig. 8] The validation is performed on only five held-out experiments (75 training), although the 15-replicate shuffled procedure adds some robustness. The reported prediction errors 'within ±5 mm near the sheet tip' are large relative to the 1.5 mm control tolerance; the paper should comment on this mismatch and explain why the surrogate is nevertheless adequate for MPC.
  4. [§5.1, Eq. (20)] The constraint h2: c1^T ũ_k = 0 and g2: c2^T ũ_k < 0 are used to enforce toolpath endpoint behavior. The derivation relies on Chebyshev properties but is not explicitly shown; a brief explanation or reference would improve readability.
  5. [References] Reference [15] appears to contain a page number '6650675' rather than a standard article/page range, and reference [13] lists the year as 2028 for a CIRP Annals article. Please verify all bibliographic entries.

Circularity Check

0 steps flagged

No significant circularity; Koopman/POD model is identified from data, validated on held-out experiments, and closed-loop outcomes are physical.

full rationale

The derivation is not circular. The Koopman/POD surrogate is identified from 75 training experiments and its one-step and multi-step predictions are checked against five held-out experiments (Sec. 4.4, Figs. 8b–e); this is standard system identification with independent validation, not a fit renamed as a prediction. The MPC objective (Eq. 20) minimizes deviation from a user-specified target r, and the RLS update (Eqs. 21–24) adjusts B from measured post-cycle geometry to correct model mismatch; these are genuine closed-loop feedback operations, and the final shapes in Sec. 5.3 are physical measured outcomes. The many self-citations (e.g., Refs. 5, 6, 7, 15, 16) concern the prior robotic English wheel hardware, metrology, and earlier DT/MPC formulations; none is invoked as an external theorem that forces the present model choice. The paper itself acknowledges the first-order Markov/material-state limitation in Sec. 5.2 ('Given the limited training data, material state was not included as an additional input'), which is a modeling weakness rather than circularity. The abstract's unqualified 'achieve the given target shape' is stronger than the reported large-deformation result (2.79 mm final max deviation vs. the 1.5 mm termination criterion), but that is an evidence-to-claim gap, not a definitional/fitted-input circularity.

Axiom & Free-Parameter Ledger

8 free parameters · 6 axioms · 0 invented entities

The framework rests on established methods (PCA, Chebyshev, Koopman, MPC, RLS) and introduces no new physical entities. The main assumptions are that the process is Markovian in the measured shape, that a finite Koopman model can capture the nonlinear dynamics, and that the PCA/Chebyshev bases cover the operating and target space. Several tuning parameters (loss weights, MPC penalties, RLS factors, update triggers) are hand-chosen; none are fitted to the specific closed-loop targets, but they do affect the results and should be disclosed in any reproduction.

free parameters (8)
  • Number of retained PCA modes r = 4
    Chosen because the first four modes capture >99.5% variance (Section 4.2); affects model state dimension and generalization.
  • Number of Chebyshev basis functions p = 5
    Chosen to represent smooth toolpaths with reconstruction error mostly within ±0.25 mm (Section 4.2).
  • Lifted dimension d_z = 256 + 4 = 260
    Architecture choice in Table 1; 256 learned observables plus 4 augmented reduced-order states.
  • Loss weights α1..α5 and λ = 1000, 20, 10, 1000, 5, 0.01
    Tuned heuristically (Section 4.3); shape the learned Koopman model but are not fit to the closed-loop targets.
  • MPC weights Q and R = Q has 1s with last four 20,10,10,1; R=1e-5
    Hand-set in Section 5.1 to prioritize tracking of the physical reduced states over observables; not learned from data.
  • RLS forgetting factor λ and initial covariance = λ=0.9, P=diag(0.001)
    Chosen in Section 5.2; controls how quickly B is adapted.
  • RLS update triggers = prediction deviation >3 mm or stagnation <3 mm
    Hand-set in Section 5.2; decides when online adaptation is applied.
  • Prediction horizon N and rollout length S_p = N=6, S_p=4
    Fixed in Sections 4.3 and 5.1; affect MPC optimization size and multi-step training loss.
axioms (6)
  • domain assumption The forming process is first-order Markovian in the measured deformation state (x_{k+1}=f(x_k,u_k)) with no explicit material-state variable.
    Section 2 defines the state as post-cycle deformation only; Section 5.2 explicitly states material state was not included as an input. Strain history is known to affect plastic response.
  • domain assumption A finite-dimensional Koopman invariant subspace exists and is learnable by the encoder–decoder network, i.e., z_{k+1}=A z_k + B u_k accurately approximates the true dynamics.
    Section 4.3 adopts the deep Koopman framework [32]; this is a standard but non-trivial assumption for nonlinear systems.
  • domain assumption The PCA basis computed from training deformation fields spans the target geometries and all new process states encountered in closed-loop control.
    Section 5.1 uses z_r=ψ(r) without verifying that the target geometry r lies in the span of the training PCA modes or within the Koopman model's validity region.
  • domain assumption Toolpaths generated by the 5-coefficient Chebyshev representation cover the executable and useful input space, and the Chebyshev constraints h2 and g2 correctly enforce physical boundary conditions.
    Section 4.2 and 5.1 rely on Chebyshev polynomial properties to guarantee smoothness and endpoint constraints; this presumes the five retained coefficients are sufficient.
  • ad hoc to paper Correcting only the B matrix (while fixing A, encoder, and decoder) is sufficient to compensate for model mismatch and material drift.
    Section 5.2 assumes "the inaccurate B is the only source that contributes to ε_k" (Eq. 21). This is a deliberate simplification, not validated against alternatives.
  • domain assumption The eight-marker midline deformation profile adequately represents the sheet state for control purposes.
    Section 2 defines the state with n_x=8 Vicon trackers and averages edge markers; out-of-plane displacement along the mid-line is assumed sufficient.

pith-pipeline@v1.3.0-alltime-deepseek · 16362 in / 13375 out tokens · 122023 ms · 2026-08-03T22:20:23.317407+00:00 · methodology

0 comments
read the original abstract

Digital Twin (DT) technologies are transforming manufacturing by enabling real-time prediction, monitoring, and control of complex processes. Yet, applying DT to deformation-based metal forming remains challenging because of the strongly coupled spatial-temporal behavior and the nonlinear relationship between toolpath and material response. For instance, sheet-metal forming by the English wheel, a highly flexible but artisan-dependent process, still lacks digital counterparts that can autonomously plan and adapt forming strategies. This study presents an adaptive DT framework that integrates Proper Orthogonal Decomposition (POD) for physics-aware dimensionality reduction with a Koopman operator for representing nonlinear system in a linear lifted space for the real-time decision-making via model predictive control (MPC). To accommodate evolving process conditions or material states, an online Recursive Least Squares (RLS) algorithm is introduced to update the operator coefficients in real time, enabling continuous adaptation of the DT model as new deformation data become available. The proposed framework is experimentally demonstrated on a robotic English Wheel sheet metal forming system, where deformation fields are measured and modeled under varying toolpaths. Results show that the adaptive DT is capable of controlling the forming process to achieve the given target shape by effectively capturing non-stationary process behaviors. Beyond this case study, the proposed framework establishes a generalizable approach for interpretable, adaptive, and computationally-efficient DT of nonlinear manufacturing systems, bridging reduced-order physics representations with data-driven adaptability to support autonomous process control and optimization.

Figures

Figures reproduced from arXiv: 2511.10852 by Derick Suarez, Guanzhong Hu, Jian Cao, Ping Guo, Vispi Karkaria, Wei Chen, Ying-Kuan Tsai, Yi-Ping Chen, Zihan Chen.

Figure 1
Figure 1. Figure 1: (a) The proposed framework for the dynamic process DT. (b) System setup for the robotic English wheel. [PITH_FULL_IMAGE:figures/full_fig_p002_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: (a) Vicon tracker locations on the blank sheet, the red circle indicates the contact point with the wheel at the [PITH_FULL_IMAGE:figures/full_fig_p004_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: Illustration of Koopman operator theory. It shows that the evolution in the state space using nonlinear model [PITH_FULL_IMAGE:figures/full_fig_p005_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: First six Chebyshev polynomials of the first kind, [PITH_FULL_IMAGE:figures/full_fig_p007_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: Dimensional reduction of sheet deformation using PCA. (a) The eigenvalue associated with each principal [PITH_FULL_IMAGE:figures/full_fig_p010_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: Dimensional reduction of toolpaths using Chebyshev polynomials. (a) Comparison between the reconstructed [PITH_FULL_IMAGE:figures/full_fig_p011_6.png] view at source ↗
Figure 7
Figure 7. Figure 7: The proposed proper orthogonal decomposition (POD)-based deep Koopman operator for sheet metal forming. [PITH_FULL_IMAGE:figures/full_fig_p011_7.png] view at source ↗
Figure 8
Figure 8. Figure 8: Validation of the learned Koopman operator. (a) Training and validation loss along epochs. (b) and (c) [PITH_FULL_IMAGE:figures/full_fig_p013_8.png] view at source ↗
Figure 9
Figure 9. Figure 9: Result of sheet deformation over six cycles with model updating. (a) Sheet deformations over cycles. (b) Applied [PITH_FULL_IMAGE:figures/full_fig_p016_9.png] view at source ↗
Figure 10
Figure 10. Figure 10: Result of sheet deformation over six cycles without model updating. (a) Sheet deformations over cycles. (b) [PITH_FULL_IMAGE:figures/full_fig_p016_10.png] view at source ↗
Figure 11
Figure 11. Figure 11: Result of sheet deformation given moderate-deformation target over five cycles without model updating. (a) [PITH_FULL_IMAGE:figures/full_fig_p017_11.png] view at source ↗
Figure 12
Figure 12. Figure 12: The increment of B matrix during model updating from one iteration. 6 Closure In this work, we present an adaptive DT framework that integrates an adaptive POD-based deep Koopman operator for identifying the spatio-temporal dynamics of sheet deformation under toolpath actuation, followed by MPC for decision￾making. Experimental validation on a robotic English Wheel system demonstrated that the proposed fr… view at source ↗

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score.

  1. A Continual Validation, Updating, and Decision-Making Framework for Self-Adaptive Digital Twins via Robust Model Predictive Control: A Case Study in Additive Manufacturing

    cs.LG 2026-07 conditional novelty 6.0

    An adaptive digital twin framework integrating Fisher-score drift detection, LoRA fine-tuning, and Mann-Whitney U validation restores predictive accuracy and uncertainty calibration under concept drift.

Reference graph

Works this paper leans on

41 extracted references · 1 linked inside Pith · cited by 1 Pith paper

  1. [1]

    An optimization-centric review on integrating artificial intelligence and digital twin technologies in manufacturing

    Karkaria, V., Tsai, Y.-K., Chen, Y.-P., and Chen, W., 2024. “An optimization-centric review on integrating artificial intelligence and digital twin technologies in manufacturing”.Engineering Optimization, pp. 1–47

    2024

  2. [2]

    A probabilistic graphical model foundation for enabling predictive digital twins at scale

    Kapteyn, M. G., Pretorius, J. V., and Willcox, K. E., 2021. “A probabilistic graphical model foundation for enabling predictive digital twins at scale”.Nature Computational Science,1(5), pp. 337–347

    2021

  3. [3]

    Energy digital twin technology for industrial energy management: Classification, challenges and future

    Yu, W., Patros, P., Young, B., Klinac, E., and Walmsley, T. G., 2022. “Energy digital twin technology for industrial energy management: Classification, challenges and future”.Renewable and Sustainable Energy Reviews,161, p. 112407

    2022

  4. [4]

    Leveragingaiforenergy-efficientmanufacturing systems: Review and future prospectives

    Abadi,M.M.K.F.,Liu,C.,Zhang,M.,Hu,Y.,andXu,Y.,2025. “Leveragingaiforenergy-efficientmanufacturing systems: Review and future prospectives”.Journal of Manufacturing Systems,78, pp. 153–177

    2025

  5. [5]

    Towards a digital twin framework in additive manufacturing: Machine learning and bayesian optimization for time series process optimization

    Karkaria,V.,Goeckner,A.,Zha,R.,Chen,J.,Zhang,J.,Zhu,Q.,Cao,J.,Gao,R.X.,andChen,W.,2024. “Towards a digital twin framework in additive manufacturing: Machine learning and bayesian optimization for time series process optimization”.Journal of Manufacturing Systems

    2024

  6. [6]

    Real-time decision-making for digital twin in additive manufacturing with model predictive control using time-series deep neural networks

    Chen, Y.-P., Karkaria, V., Tsai, Y.-K., Rolark, F., Quispe, D., Gao, R. X., Cao, J., and Chen, W., 2025. “Real-time decision-making for digital twin in additive manufacturing with model predictive control using time-series deep neural networks”.Journal of Manufacturing Systems,80, pp. 412–424

    2025

  7. [7]

    Uncertainty-awaredigitaltwins: Robustmodelpredictive control using time-series deep quantile learning

    Chen,Y.-P.,Tsai,Y.-K.,Karkaria,V.,andChen,W.,2026.“Uncertainty-awaredigitaltwins: Robustmodelpredictive control using time-series deep quantile learning”.Journal of Mechanical Design,148(2), p. 021702

    2026

  8. [8]

    Digitaltwin-enabledmulti-generationcontrolco-design with deep reinforcement learning

    Tsai,Y.-K.,Karkaria,V.,Chen,Y.-P.,andChen,W.,2025. “Digitaltwin-enabledmulti-generationcontrolco-design with deep reinforcement learning”.arXiv preprint arXiv:2510.10694

    arXiv 2025

  9. [9]

    Machine learning-assistedin-situadaptivestrategiesforthecontrolofdefectsandanomaliesinmetaladditivemanufacturing

    Gunasegaram,D.,Barnard,A.,Matthews,M.,Jared,B.,Andreaco,A.,Bartsch,K.,andMurphy,A.,2024. “Machine learning-assistedin-situadaptivestrategiesforthecontrolofdefectsandanomaliesinmetaladditivemanufacturing”. Additive Manufacturing, p. 104013

    2024

  10. [10]

    An attention-based spatio-temporal neural operator for evolving physics

    Karkaria, V., Lee, D., Chen, Y.-P., Yu, Y., and Chen, W., 2025. “An attention-based spatio-temporal neural operator for evolving physics”.Machine Learning: Science and Technology

    2025

  11. [11]

    Sheet metal market size, share & growth report, 2035.https://www

    Market Research Future, 2025. Sheet metal market size, share & growth report, 2035.https://www. marketresearchfuture.com/reports/sheet-metal-market-8584. Accessed: 2025-10-28

    2025

  12. [12]

    From art to part: Learning from the traditional smith in developing flexible sheet metal forming processes

    Bowen, D. T., Russo, I. M., Cleaver, C. J., Allwood, J. M., and Loukaides, E. G., 2022. “From art to part: Learning from the traditional smith in developing flexible sheet metal forming processes”.Journal of Materials Processing Technology,299, p. 117337

    2022

  13. [13]

    Flexibility in metal forming

    Yang, D., 2028. “Flexibility in metal forming”.CIRP Annals,67(2), pp. 743–765

    2028

  14. [14]

    A forging technology first: Integrating industry 4.0 into an autonomous cnc forging machine

    Bandar, A., Thurston, B., Daehn, G., Groeber, M., Wright, C., Bianco, N., and Trzinski, S., 2025. “A forging technology first: Integrating industry 4.0 into an autonomous cnc forging machine”.FIA magazine,7. 18

    2025

  15. [15]

    On the feasability of an integrated english wheel system

    Suarez, D., Chen, F., Kang, P., Forbes, B., Gao, M., Ineza, O., Benton, K., and Dewberry, N., 2024. “On the feasability of an integrated english wheel system”.Journal of Manufacturing Systems,75(9), p. 6650675

    2024

  16. [16]

    In-process part tracking and shape measurement using vision-based motion capture for automated english wheeling

    Zhang, Y., Suarez, D., Ehmann, K., Cao, J., and Guo, P., 2024. “In-process part tracking and shape measurement using vision-based motion capture for automated english wheeling”.Manufacturing Letters,41, pp. 241–247

    2024

  17. [17]

    Closed-loop control of product properties in metal forming

    Allwood,J.,Duncan,S.,Cao,J.,Groche,P.,Hirt,G.,Kinsey,B.,Kuboki,T.,Liewald,M.,Sterzing,A.,andTekkaya, A., 2016. “Closed-loop control of product properties in metal forming”.CIRP Annals,65(2), pp. 573–596

    2016

  18. [18]

    Reinforcement learning in free-form stamping of sheet-metals

    Liu, S., Shi, Z., Lin, J., and Li, Z., 2020. “Reinforcement learning in free-form stamping of sheet-metals”.Procedia Manufacturing,50, pp. 444–449

    2020

  19. [19]

    Areviewofcharacterization and modelling approaches for sheet metal forming of lightweight metallic materials

    Hou,Y.,Myung,D.,Park,J.K.,Min,J.,Lee,H.-R.,El-Aty,A.A.,andLee,M.-G.,2023.“Areviewofcharacterization and modelling approaches for sheet metal forming of lightweight metallic materials”.Materials,16(2), p. 836

    2023

  20. [20]

    Model predictive control of incre- mental sheet forming for geometric accuracy improvement

    Lu, H., Kearney, M., Li, Y., Liu, S., Daniel, W. J., and Meehan, P. A., 2016. “Model predictive control of incre- mental sheet forming for geometric accuracy improvement”.The International Journal of Advanced Manufacturing Technology,82(9), pp. 1781–1794

    2016

  21. [21]

    Learning-based model predictive control for two-point incremental sheet forming

    Wang, C., He, A., Liu, S., and Meehan, P. A., 2023. “Learning-based model predictive control for two-point incremental sheet forming”.Journal of Manufacturing Processes,105, pp. 187–198

    2023

  22. [22]

    Switched model predictive path control of incremental sheet forming for parts with varying wall angles

    He, A., Wang, C., Liu, S., and Meehan, P. A., 2020. “Switched model predictive path control of incremental sheet forming for parts with varying wall angles”.Journal of Manufacturing Processes,53, pp. 342–355

    2020

  23. [23]

    Koopmaninvariantsubspacesandfinitelinear representations of nonlinear dynamical systems for control

    Brunton, S.L., Brunton, B.W., Proctor, J.L., andKutz, J.N., 2016. “Koopmaninvariantsubspacesandfinitelinear representations of nonlinear dynamical systems for control”.PloS one,11(2), p. e0150171

    2016

  24. [24]

    Notesonkoopmanoperatortheory

    Brunton,S.L.,2019. “Notesonkoopmanoperatortheory”.UniversitätvonWashington,DepartmentofMechanical Engineering, Zugriff,30

    2019

  25. [25]

    Active learning of dynamics for data-driven control using koopman operators

    Abraham, I., and Murphey, T. D., 2019. “Active learning of dynamics for data-driven control using koopman operators”.IEEE Transactions on Robotics,35(5), pp. 1071–1083

    2019

  26. [26]

    Data-drivencontrolofsoftrobotsusing koopman operator theory

    Bruder,D.,Fu,X.,Gillespie,R.B.,Remy,C.D.,andVasudevan,R.,2020. “Data-drivencontrolofsoftrobotsusing koopman operator theory”.IEEE transactions on robotics,37(3), pp. 948–961

    2020

  27. [27]

    Retrofit: Real-time control of time-dependent 3d point cloud profiles

    Biehler, M., and Shi, J., 2024. “Retrofit: Real-time control of time-dependent 3d point cloud profiles”.Journal of Manufacturing Science and Engineering,146(6), p. 061006

    2024

  28. [28]

    Efficient approximation of the koopman operator for large-scale nonlinear systems

    Verma, G., Heath, W., and Theodoropoulos, C., 2025. “Efficient approximation of the koopman operator for large-scale nonlinear systems”. In ESCAPE 35-European Symposium on Computer Aided Process Engineering, pp. 1251–1256

    2025

  29. [29]

    Analysisoftime-resolvedpivmeasurementsofaconfinedturbulent jet using pod and koopman modes

    Semeraro,O.,Bellani,G.,andLundell,F.,2012. “Analysisoftime-resolvedpivmeasurementsofaconfinedturbulent jet using pod and koopman modes”.Experiments in fluids,53(5), pp. 1203–1220

    2012

  30. [30]

    Mauroy, A., Susuki, Y., and Mezic, I., 2020.Koopman operator in systems and control, Vol. 7. Springer

    2020

  31. [31]

    Extended dynamic mode decompositionwithlearnedkoopmaneigenfunctionsforpredictionandcontrol

    Folkestad, C., Pastor, D., Mezic, I., Mohr, R., Fonoberova, M., and Burdick, J., 2020. “Extended dynamic mode decompositionwithlearnedkoopmaneigenfunctionsforpredictionandcontrol”.In2020americancontrolconference (acc), IEEE, pp. 3906–3913

    2020

  32. [32]

    Deep learning for universal linear embeddings of nonlinear dynamics

    Lusch, B., Kutz, J. N., and Brunton, S. L., 2018. “Deep learning for universal linear embeddings of nonlinear dynamics”.Nature communications,9(1), p. 4950

    2018

  33. [33]

    Robust control co-design using tube-based model predictive control

    Tsai, Y.-K., and Malak, R. J., 2023. “Robust control co-design using tube-based model predictive control”. In 2023 American Control Conference (ACC), IEEE, pp. 769–775. 19

    2023

  34. [34]

    Reduced-orderadaptivecontrollersforfluidflowsusingpod

    Ravindran,S.,2000. “Reduced-orderadaptivecontrollersforfluidflowsusingpod”.Journalofscientificcomputing, 15(4), pp. 457–478

    2000

  35. [35]

    Unknown parameter excitation and estimation for complex systems with dynamic performances

    Chen, Y.-P., and Chan, K.-Y., 2021. “Unknown parameter excitation and estimation for complex systems with dynamic performances”.Journal of Mechanical Design,143(9), p. 091704

    2021

  36. [36]

    C., and Handscomb, D

    Mason, J. C., and Handscomb, D. C., 2002.Chebyshev polynomials. Chapman and Hall/CRC

    2002

  37. [37]

    Inception-v4,inception-resnetandtheimpactofresidual connections on learning

    Szegedy,C.,Ioffe,S.,Vanhoucke,V.,andAlemi,A.,2017. “Inception-v4,inception-resnetandtheimpactofresidual connections on learning”. In Proceedings of the AAAI conference on artificial intelligence, Vol. 31

    2017

  38. [38]

    OSQP: an operator splitting solver for quadratic programs

    Stellato, B., Banjac, G., Goulart, P., Bemporad, A., and Boyd, S., 2020. “OSQP: an operator splitting solver for quadratic programs”.Mathematical Programming Computation,12(4), pp. 637–672

    2020

  39. [39]

    Koopman operator-based model predictive control with recursive online update

    Calderón, H. M., Schulz, E., Oehlschlägel, T., and Werner, H., 2021. “Koopman operator-based model predictive control with recursive online update”. In 2021 European Control Conference (ECC), IEEE, pp. 1543–1549

    2021

  40. [40]

    Local koopman operators for data-driven control of robotic systems

    Mamakoukas, G., Castano, M., Tan, X., and Murphey, T., 2019. “Local koopman operators for data-driven control of robotic systems”. In Robotics: science and systems

    2019

  41. [41]

    Model predictive control of nonlinear dynamics using online adaptive koopman operators

    Uchida, D., and Duraisamy, K., 2024. “Model predictive control of nonlinear dynamics using online adaptive koopman operators”.arXiv:2412.02972. 20

    Pith/arXiv arXiv 2024