Pith. sign in

REVIEW 5 major objections 8 minor 43 references

A Novel Feedforward Youla Parameterization Method for Avoiding Local Minima in Stereo Image Based Visual Servoing Control

T0 review · 5 major / 8 minor · reviewed 2026-08-07 · deepseek-v4-flash

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

Pith's one-line read Three non-collinear points fix a stereo camera pose, and a feedforward Youla feedback loop lets stereo IBVS match all nine image coordinates instead of falling into local minima.

desk verdict The feedforward + Youla idea is plausible, but the P3P 'proof' omits the world-frame match and the simulation comparison handicaps the feedback-only controller by design, so the central claims don't hold as written. read the letter →

arxiv 2506.10252 v1 pith:MYGS2FYO submitted 2025-06-12 cs.RO cs.SYeess.SY

Rongfei Li , Francis Assadian This is my paper · ORCID
classification cs.ROcs.SYeess.SY MSC 93C8593B52
keywords stereoimage-basedvisualservoingP3PproblemPnPlocalminimafeedforwardcontrolYoulaparameterizationeye-in-handconfigurationcameraposeestimation
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 aims to remove local minima from stereo image-based visual servoing by changing how the control problem is posed. It first argues that a calibrated stereo camera's pose can be uniquely determined from three known non-collinear points, since stereo triangulation gives the three-dimensional coordinates of each point in the camera frame and three points fix a unique frame. Because each point yields three image coordinates, three markers create nine constraints on the six-degree-of-freedom camera motion, and the paper shows in simulation that a feedback-only IBVS controller matches only six of those constraints and gets trapped. The proposed solution is a joint-space feedforward controller that commands the desired joint configuration directly, combined with an adaptive Youla-parameterized feedback loop for steady tracking. In simulation the combined controller makes all nine image coordinates match the target with and without a one-degree joint disturbance, which matters because it offers a local-minima-free stereo IBVS scheme using linear control design rather than mode switching or per-step optimization.

What carries the argument

The carrying mechanism is the feedforward-feedback cascade in joint space. The inner loop feedback-linearizes the six-link robot to $\ddot{q} = v$ and closes it with a Youla-parameterized controller whose closed-loop transfer function is $T_{\mathrm{inner}}=(3\tau_{\mathrm{in}}s+1)/(\tau_{\mathrm{in}}s+1)^3$ per joint. The feedforward path commands $q_{\mathrm{desired}}$ through approximately the inverse of this response, $T_{\mathrm{ff}}=((\tau_{\mathrm{in}}s+1)^3/(3\tau_{\mathrm{in}}s+1))\cdot 1/(\tau_{\mathrm{ff}}s+1)^2$ with $\tau_{\mathrm{ff}}=0.1\tau_{\mathrm{in}}$, so the robot moves quickly toward the target pose without using image features. The outer loop linearizes the eye-in-hand map $p=\mathcal{F}(q)$ about the current estimated joint state as $p=C_1 q+C_2$, diagonalizes $C_1$ through singular value decomposition, and applies a Youla controller shaped by the closed-loop target $M_T$, whose last $3\times 3$ block is zero, meaning the third marker's three coordinates are deliberately left to the feedforward path. The geometric basis is the paper's proof that three non-collinear points fix a unique stereo camera frame through the plane normal $n=(P_1P_2\times P_1P_3)/\|P_1P_2\times P_1P_3\|$, which is why nine image coordinates are both necessary and sufficient and why six feedback outputs cannot resolve all nine constraints.

What would settle it

Generate two distinct camera poses viewing three non-collinear points and check whether both poses give the same three stereo-triangulated coordinates; if they do, the P3P sufficiency claim is false. Separately, add a deliberate 1-degree calibration bias to the inverse kinematics that produces $q_{\mathrm{desired}}$ and check whether the third marker's three image coordinates still converge, since feedback leaves them uncontrolled.

Watch

Extended reading notes

Core claim

The paper establishes two linked claims. First, for a calibrated stereo camera, the perspective-n-point problem is solved by exactly three points: because stereo triangulation recovers metric 3D coordinates in the camera frame, three non-collinear points determine the plane normal and hence the camera frame's axes and origin uniquely, so three markers, giving nine image coordinates, determine the six-degree-of-freedom pose. Second, this overdetermination is what makes feedback-only IBVS fail: with nine image constraints and only six joint degrees of freedom, the best a feedback-only controller can do in the simulations is match six of the nine coordinates, leaving the third marker's coordinates stuck away from target. The proposed controller adds a feedforward path that commands the joint configuration corresponding to the desired camera pose, while an adaptively re-linearized Youla-parameterized feedback loop handles disturbance rejection and fine tracking of six coordinates. In the two simulated scenarios, one without disturbances and one with a 1 degree step disturbance on every joint, the combined controller makes all nine image coordinates converge to their target values, whereas the feedback-only controller does not.

Load-bearing premise

The scheme assumes the controller knows the exact joint configuration that puts the camera at the target pose and that the robot can track it, because the feedback loop deliberately leaves the third marker's three image coordinates uncontrolled, so any error in that feedforward command cannot be corrected.

Editorial extensions

If this is right

  • A calibrated stereo IBVS system needs only three fiducial markers; the fourth correspondence is redundant for pose determination.
  • A feedback-only stereo IBVS controller will, at best, match six of the nine image coordinates, because nine constraints exceed the six controlled degrees of freedom.
  • Adding the feedforward joint-space command makes all nine target image coordinates reachable in the simulated scenarios, with and without a 1-degree input disturbance.
  • The combined controller converges in under two seconds in the simulations, compared with under three seconds for feedback-only control, at the price of possible overshoot from the feedforward action.
  • The architecture avoids local minima without switching control modes, keeping the pure IBVS structure and avoiding per-step optimization.

Reading between the lines

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

  • A testable extension would be to use the third marker's three residual coordinates as an outer correction signal for $q_{\mathrm{desired}}$, restoring what-to-do if the feedforward model is imperfect.
  • The P3P argument effectively turns stereo IBVS into a 3D-3D registration problem, so a direct point-cloud alignment loop may replace the image-space Jacobian and its linearization about a nominal configuration.
  • If the feedforward command is what avoids local minima, the controller's basin of attraction should be measured from many initial poses; the paper reports only two scenarios.
  • The same feedforward strategy could apply to other overdetermined visual servoing tasks by assigning every constraint beyond the six controlled degrees of freedom to the feedforward path, at the cost of trusting calibration.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

5 major / 8 minor

Summary. The paper makes two central claims. First, it claims to prove that the Perspective-3-Point (P3P) problem is sufficient to determine the pose of a calibrated stereo camera, since each point yields three image coordinates and hence three 3D coordinates in the camera frame. Second, it proposes a feedforward-plus-feedback controller for stereo image-based visual servoing (IBVS), where the feedforward path commands the desired robot joint configuration obtained from inverse kinematics and the feedback path uses a Youla-parameterized adaptive controller. The authors compare feedback-only and feedforward-feedback controllers in two simulation scenarios (with and without a 1° joint input disturbance) and conclude that the feedforward-feedback scheme avoids local minima and matches all nine image coordinates, while the feedback-only scheme does not.

Significance. The underlying observation that three points with known depths determine the camera pose is correct in substance, and the two-loop Youla parameterization is a coherent classical control design. If the local-minima-avoidance claim were established, the architecture could be a useful practical contribution. However, the paper's control claim rests on a heavily biased comparison: the feedback loop deliberately leaves three of the nine image coordinates uncontrolled, so the feedback-only simulator is not a standard 9-output IBVS baseline. In addition, the P3P proof in Section 3 is incomplete as written, and the load-bearing exactness of the feedforward command is never tested against calibration or inverse-kinematics errors. The paper contains a correct Kabsch-style rigid-alignment derivation in Appendix C, but it is not connected to the Section 3 proof, and no code or reproducibility artifacts are provided. The nominal simulations are clean, but they do not support the paper's general claims.

major comments (5)
  1. [Section 3, Eqs. (1)-(4)] The P3P proof is incomplete and, as written, circular. The argument uses the plane normal n and a sphere centered at P1 to argue that the camera coordinate system is uniquely determined, but it never uses the known world-frame coordinates of the three points, which are the data that fix the rotation between the camera frame and the inertial frame. Three non-collinear points define a plane, but they do not by themselves determine the orientation of the camera frame; one also needs the correspondence between the coordinates of those points in the camera frame and in the world frame. The correct derivation is essentially the Kabsch-style alignment in Appendix C, but the paper does not link that derivation to the Section 3 conclusion. The assertion "In conclusion, the P3P problem is sufficient to solve the PnP problem for a stereo camera system" is therefore not supported by the proof actually presented.
  2. [Section 6.3, Eq. (56) and Figures 14/16] The feedback-only comparison does not test local-minima avoidance. Equation (56) defines the desired closed-loop map M_T with a 3x3 zero block, and the text explicitly states "The last 3 coordinates remain uncontrolled in the feedback loop." Consequently, the feedback-only controller is designed not to control the third marker's image coordinates. The steady-state mismatch of those three coordinates in Figures 14 and 16 is therefore a consequence of the control design, not of the system being trapped in a local minimum. To support the paper's central claim, the baseline would need to be a genuine 9-output IBVS controller (e.g., one using a least-squares pseudo-inverse of the 9x6 interaction matrix) whose failure to match all features can then be attributed to local minima or overdetermination.
  3. [Section 7, Scenario 2 and Eq. (48)] The robustness claim is not tested against the principal failure mode of the proposed architecture. The feedforward path assumes that the desired joint configuration q_desired computed from inverse kinematics and hand-eye calibration is exactly correct. Because the feedback loop leaves the last three image coordinates uncontrolled, any error in q_desired arising from inverse-kinematics approximation, robot kinematic parameter error, or hand-eye calibration error T_E^C will appear directly as a steady-state error in the third marker's coordinates, with no feedback correction possible. The 1° step input disturbance in Scenario 2 is rejected by the inner loop and does not exercise this failure mode. No sensitivity analysis over q_desired or T_E^C errors is provided, so the claim that the method is robust to disturbances and model uncertainties is unsupported.
  4. [Section 6.3, Eqs. (51) and (53)] The Jacobian dimensions in Section 6.3 are inconsistent. The mapping F maps R^6 (joint angles) to R^9 (image coordinates), so its Jacobian must be 9x6. However, the text states J(q0, P_T, P_O) ∈ R^{6x6} and C1 ∈ R^{6x6}, while the subsequent SVD decomposition in Eq. (54) correctly uses U_L ∈ R^{9x9}, U_R ∈ R^{6x6}, and M_p ∈ R^{9x6}. The dimension statements need to be corrected, and the definitions of C1 and the plant transfer function in Eq. (53) must be made consistent with a 9x6 or 9x1 output structure.
  5. [Section 4.2, Eq. (29)] The robot dynamic model in Eq. (29) has dimensional inconsistencies. The paper declares B ∈ R^{6x1} and T ∈ R^{6x1}, but then writes (C(q,q_dot) + B^T) q_dot, which is not dimensionally consistent unless B and T are matrices or the notation is intended differently. Since the inner-loop design and feedback linearization in Section 6.1 rest on this model, the notation should be corrected or clarified.
minor comments (8)
  1. [Keywords] The keyword "Sterero camera system" contains a typo and should read "Stereo camera system."
  2. [Section 3] "Prove concluded" should read "Proof concluded."
  3. [Section 7] The sentence "Amon 9g these denoising methods" contains a typo and should read "Among these denoising methods."
  4. [Section 4.6, Eq. (26)] The matrices in Eqs. (26) and (27) are typeset with misaligned columns; the intended 4x4 homogeneous transformation format is not clear from the printed formulas.
  5. [Appendix B] The notation c_{i,j} and s_{i,j} is not fully defined: Eq. (B5) defines c_{i,j} but omits the definition of s_{i,j}.
  6. [References] References [33] and [37] are anonymous product manuals with only an access date; this is insufficient citation practice for a journal submission.
  7. [Simulation Setup] The table of initial and final poses in Section 7 is garbled and difficult to parse; please reformat it so the entries for each scenario are clearly aligned.
  8. [General] No code, Simulink model, or reproducibility data is provided, which makes the two simulated scenarios impossible to verify independently.

Circularity Check

2 steps flagged · score 4.0 of 10

Feedback-only 'local minimum' is designed in via Eq. (56); the all-nine-coordinate success of the feedforward case largely restates that the exact target joint configuration was commanded, so the validation is partially self-fulfilling even though the P3P reduction and Youla derivations are not circular.

  1. other [Section 6.3, Eq. (56); Section 7, Figs. 14 and 16]
    "M_T = … [ I_{6×6} Z_{3×3}; Z_{3×3} Z_{3×3} ] … The last 3 coordinates remain uncontrolled in the feedback loop. … the feedback-only controller fails to guide the camera to its desired pose and falls into local minima, as evident from the third point’s coordinates … which do not match the target at steady state."

    The feedback-only controller is constructed with the lower-right 3x3 block of M_T identically zero, so the feedback path cannot act on the third marker's three image coordinates. The simulation's observation that those coordinates remain unmatched is therefore fixed by the design of Eq. (56) rather than being an emergent instance of an overdetermined system falling into a local minimum. Presenting that designed-in outcome as evidence for the local-minima claim makes the demonstration circular: the 'predicted' failure is an input of the controller definition.

  2. self definitional [Section 7, feedforward discussion after Figs. 15 and 16]
    "This is because the feedforward component directly controls the robot’s joint angles rather than image features. Since the joint angles (6 DoFs) uniquely correspond to the camera’s pose (6 DoFs), the feedforward controller helps the system reach the global minimum by using the desired joint configurations as inputs."

    The feedforward command is q_desired, the inverse-kinematic solution for the target camera pose, and the simulated plant is the same kinematic/dynamic model used to compute it. Commanding the target configuration trivially drives the camera to the target pose when the model and calibration are exact, so the matching of all nine image coordinates is guaranteed by construction rather than being evidence of a general local-minima-avoidance mechanism. The claimed advantage over feedback-only control therefore rests on the exactness of the feedforward input, which is assumed, not derived.

full rationale

No fitted parameters are presented as predictions, and the Youla controller derivations are self-contained, so the paper is not circular in the sense of fitting data or leaning on a self-citation chain. The P3P-to-stereo reduction is an elementary but valid observation: metric stereo triangulation gives 3D points in the camera frame, and three non-collinear correspondences determine a rigid pose. The main circularity lies in the validation logic: the feedback-only controller is deliberately designed with the last three image coordinates uncontrollable (Eq. 56), so its failure to match those coordinates is by construction; and the feedforward controller's success is essentially the tautology of commanding the exact desired joint configuration in an exact model. These two steps make the simulation comparison partially self-fulfilling, although the proposed control architecture itself retains independent design content.

Assumptions & free parameters 4 free parameters · 5 assumptions · 0 invented entities

The paper introduces no new physical entities. Its central claims rest on standard rigid-body geometry, exact stereo triangulation, exact feedback linearization, and the availability of a desired joint configuration through inverse kinematics. The controller tuning parameters are hand-chosen but do not represent fits to measured data.

free parameters (4)
  • inner loop time constant tau_in = 0.01 s
    Set by assuming a 100 rad/s inner-loop bandwidth for industrial actuators (Section 7, Eq. 63).
  • outer loop natural frequency omega_outer = 10 rad/s
    Chosen as one-tenth of the inner-loop bandwidth to enforce time-scale separation (Section 7, Eq. 64).
  • outer loop damping ratio zeta = 10
    Selected large to suppress overshoot in the Butterworth-shaped closed-loop response (Eq. 66).
  • feedforward filter time constant tau_ff = 0.001 s
    Placed at one-tenth of the inner-loop time constant to keep the feedforward inverse proper (Eqs. 49 and 65).
assumptions (5)
  • domain assumption Stereo triangulation recovers exact 3D camera-frame coordinates for each observed point from its two image projections (Eqs. C3-C8).
    The P3P proof and the eye-in-hand mapping M require this; Section 7 assumes image noise is almost perfectly attenuated.
  • standard math Three non-collinear 3D point correspondences determine a unique rigid transform between the camera and inertial frames.
    This is the absolute orientation result the Section 3 proof tries to establish, but the proof as written uses only the camera-frame normal and does not invoke the inertial-frame marker coordinates.
  • domain assumption The robot dynamic model in Eq. (29) is known exactly, so feedback linearization in Eq. (33) yields six decoupled double integrators.
    The inner-loop Youla design cancels the nonlinear dynamics; unmodeled effects are lumped into the input disturbance d_qT.
  • domain assumption A globally valid inverse of the eye-in-hand mapping F^{-1} exists and is computable online (Eq. 62).
    The adaptive outer loop re-estimates q_hat from image coordinates; this requires solvable inverse kinematics and a non-degenerate Jacobian throughout the trajectory.
  • domain assumption The desired joint configuration q_desired associated with the target camera pose is known and reachable.
    The feedforward controller in Eq. (48) commands q_desired; because the last three image coordinates are not controlled by feedback (Eq. 56), feedforward error cannot be corrected.

how reviews work

0 comments
Cite this review

Pith. "Pith review of A Novel Feedforward Youla Parameterization Method for Avoiding Local Minima in Stereo Image Based Visual Servoing Control." pith.science (2026). https://pith.science/paper/MYGS2FYO

@misc{pith2026250610252,
  author       = {Pith},
  title        = {Pith review of: A Novel Feedforward Youla Parameterization Method for Avoiding Local Minima in Stereo Image Based Visual Servoing Control},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/MYGS2FYO}},
  note         = {Machine review of arXiv:2506.10252}
}
read the original abstract

In robot navigation and manipulation, accurately determining the camera's pose relative to the environment is crucial for effective task execution. In this paper, we systematically prove that this problem corresponds to the Perspective-3-Point (P3P) formulation, where exactly three known 3D points and their corresponding 2D image projections are used to estimate the pose of a stereo camera. In image-based visual servoing (IBVS) control, the system becomes overdetermined, as the 6 degrees of freedom (DoF) of the stereo camera must align with 9 observed 2D features in the scene. When more constraints are imposed than available DoFs, global stability cannot be guaranteed, as the camera may become trapped in a local minimum far from the desired configuration during servoing. To address this issue, we propose a novel control strategy for accurately positioning a calibrated stereo camera. Our approach integrates a feedforward controller with a Youla parameterization-based feedback controller, ensuring robust servoing performance. Through simulations, we demonstrate that our method effectively avoids local minima and enables the camera to reach the desired pose accurately and efficiently.

Figures

Figures reproduced from arXiv: 2506.10252 by the authors.

Figure 2
Figure 2. P3P case of a monocular camera. The PnP problem with a stereo camera has not been thoroughly addressed in prior research. A stereo camera can detect three image coordinates of a 3D point in space. This paper proposes that a complete solution to the PnP problem for a stereo camera can be framed as a P3P problem. Below is the complete proof of this proposition. Proof: For a stereo system, if all intrinsic parameters a… view at source ↗
Figure 4
Figure 4. P2P problem with a stereo camera system. All potential cartesian systems are located on the circle plotted in red. 3) P3P problem with the stereo camera: When three points in space are known, and the lines connecting these points are not col￾linear, we can uniquely establish one coordinate system. As illustrated in the figure below, three non-collinear points define a plane in space, which has a uniquely defined nor… view at source ↗
Figure 5
Figure 5. P3P Problem with a Stereo Camera System. Prove concluded This proposition indicates that to uniquely determine the full 6 DoFs of the stereo camera, at least three points (or nine 2D features) are required to match in the image-based visual servoing control. 4. Model Development 4.1. Stereo Camera Model Depth between the objects to the camera plane is either approximated or estimated in the IBVS for generating the i… view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

43 extracted references · 25 canonical work pages

  1. [1]

    Improving learning -based camera pose estimation for image-based augmented reality applica- tions,

    E. Cai, R. Rossi, and C. Xiao, "Improving learning -based camera pose estimation for image-based augmented reality applica- tions," in Extended Abstracts of the 2023 CHI Conference on Human Factors in Computing Systems, Hamburg, Germany, 2023, pp. 1-

    work page 2023

  2. [2]

    Stier, N., Angles, B., Yang, L., Yan, Y., Colburn, A., & Chuang, M. (2023). LivePose: Online 3D Reconstruction from Monocular Video with Dynamic Camera Poses. In Proceedings of the IEEE/CVF International Conference on Computer Vision

    work page 2023

  3. [3]

    Hybrid Camera Pose Estimation With Online Partitioning for SLAM,

    X. Li and H. Ling, "Hybrid Camera Pose Estimation With Online Partitioning for SLAM," in IEEE Robotics and Automation Letters, vol. 5, no. 2, pp. 1453–1460, Apr. 2020. doi: 10.1109/LRA.2020.2967688

    arXiv 2020

  4. [4]

    A comparative study of pose estimation algorithms for visual navigation in autonomous robots,

    S. S. Jacob and S. S, "A comparative study of pose estimation algorithms for visual navigation in autonomous robots," Interna- tional Robotics & Automation Journal, vol. 9, no. 3, pp. 1–7, 2023. doi: 10.15406/iratj.2023.09.00343

    work page doi:10.15406/iratj.2023.09.00343 2023

  5. [5]

    Camera-to-Robot Pose Estimation from a Single Image,

    T. E. Lee, J. Tremblay, T. To, J. Cheng, T. Mosier, O. Kroemer, D. Fox, and S. Birchfield, "Camera-to-Robot Pose Estimation from a Single Image," arXiv preprint arXiv:1911.09231, 2020. [Online]. Available: https://arxiv.org/abs/1911.09231

    arXiv 1911

  6. [6]

    doi: 10.1145/3544549.3585756

    arXiv

  7. [7]

    Random Sample Consensus: A Paradigm for Model Fitting with Applications to Image Analysis and Automated Cartography

    Fischler, M. A.; Bolles, R. C. (1981). "Random Sample Consensus: A Paradigm for Model Fitting with Applications to Image Analysis and Automated Cartography". Communications of the ACM. 24 (6): 381–395. doi:10.1145/358669.358692. S2CID 972888

    arXiv 1981

  8. [8]

    A general solution to the P4P problem for camera with unknown focal length,

    M. Bujnak, Z. Kukelova and T. Pajdla, "A general solution to the P4P problem for camera with unknown focal length," 2008 IEEE Conference on Computer Vision and Pattern Recognition, Anchorage, AK, USA, 2008, pp. 1-8, doi: 10.1109/CVPR.2008.4587793

    work page arXiv 2008

Show all 43 references
  1. [9]

    EPnP: An Accurate O(n) Solution to the PnP Problem

    Lepetit, V.; Moreno-Noguer, M.; Fua, P. (2009). "EPnP: An Accurate O(n) Solution to the PnP Problem". International Journal of Computer Vision. 81 (2): 155–166. doi:10.1007/s11263-008-0152-6. hdl:2117/10327. S2CID 207252029

    2009 doi

  2. [10]

    A Consistently Fast and Globally Optimal Solution to the Perspective-n-Point Prob- lem

    Terzakis, George; Lourakis, Manolis (2020). "A Consistently Fast and Globally Optimal Solution to the Perspective-n-Point Prob- lem". Computer Vision – ECCV 2020. Lecture Notes in Computer Science. Vol. 12346. pp. 478–494. doi:10.1007/978-3-030-58452- 8_28. ISBN 978-3-030-5845...

    2020 doi

  3. [11]

    Alkhatib, A.V

    M.N. Alkhatib, A.V. Bobkov, N.M. Zadoroznaya, Camera pose estimation based on structure from motion, Procedia Computer Science, Volume 186, 2021, Pages 146-153, ISSN 1877-0509, https://doi.org/10.1016/j.procs.2021.04.205

    2021 doi

  4. [12]

    Fusion of Visual Odometry Information for Enhanced Camera Pose Estima- tion,

    J. Wang, Y. Wang, C. Guo, S. Xing and X. Ye, "Fusion of Visual Odometry Information for Enhanced Camera Pose Estima- tion," 2023 8th International Conference on Control, Robotics and Cybernetics (CRC), Changsha, China, 2024, pp. 306-309, doi: 10.1109/CRC60659.2023.10488546

    2023

  5. [13]

    Naseer T., Burgard W. Deep regression for monocular camera -based 6-dof global localization in outdoor environments 2017 IEEE/RSJ International Conference on Intelligent Robots and Systems, IROS, IEEE (2017), pp. 1525-1530

    2017

  6. [14]

    Kendall, M

    A. Kendall, M. Grimes, R. Cipolla, Posenet: A convolutional network for real-time 6-DOF camera relocalization, in: Proceedings of the IEEE International Conference on Computer Vision, 2015, pp. 2938–2946

    2015

  7. [15]

    Kendall, R

    A. Kendall, R. Cipolla, Geometric loss functions for camera pose regression with deep learning, in: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2017, pp. 5974–5983

    2017

  8. [16]

    Chaumette and S

    F. Chaumette and S. Hutchinson. Visual servo control. I. Basic approaches. IEEE Robotics & Automation Maga- zine.2006;13(4):82-90. DOI: 10.1109/MRA.2006.250573

    2006

  9. [17]

    Robotic grasping and alignment for small size components assembly based on visual servoing,

    Y. Ma, X. Liu, J. Zhang, D. Xu, D. Zhang, and W. Wu, “Robotic grasping and alignment for small size components assembly based on visual servoing,” Int. J. Adv. Manuf. Technol., vol. 106, nos. 11–12, pp. 4827–4843, Feb. 2020

    2020

  10. [18]

    Image -Based Visual Servoing for Position Alignment With Orthogonal Binocular Vision,

    T. Hao, D. Xu and F. Qin, "Image -Based Visual Servoing for Position Alignment With Orthogonal Binocular Vision," in IEEE Transactions on Instrumentation and Measurement, vol. 72, pp. 1-10, 2023, Art no. 5019010, doi: 10.1109/TIM.2023.3289560

    2023

  11. [19]

    Optimal Visual Servoing for differentially flat underactuated systems,

    M. Sheckells, G. Garimella and M. Kobilarov, "Optimal Visual Servoing for differentially flat underactuated systems," 2016 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), Daejeon, Korea (South), 2016, pp. 5541-5548, doi: 10.1109/IROS.2016.7759815

    2016

  12. [20]

    NeRF -IBVS: Visual Servo Based on NeRF for Visual Localization and Navigation

    Wang, Yuanze, et al. "NeRF -IBVS: Visual Servo Based on NeRF for Visual Localization and Navigation." Advances in Neural Information Processing Systems, 2023

    2023

  13. [21]

    Pose and Focal Length Estimation Using Two Vanishing Points with Known Camera Position

    Guo K, Cao R, Tian Y, Ji B, Dong X, Li X. Pose and Focal Length Estimation Using Two Vanishing Points with Known Camera Position. Sensors (Basel). 2023 Apr 3;23(7):3694. doi: 10.3390/s23073694. PMID: 37050754; PMCID: PMC10098530

    2023 doi

  14. [22]

    Stable Visual Servoing Through Hybrid Switched -System Control,

    N. R. Gans and S. A. Hutchinson, "Stable Visual Servoing Through Hybrid Switched -System Control," in IEEE Transactions on Robotics, vol. 23, no. 3, pp. 530-540, June 2007, doi: 10.1109/TRO.2007.895067. keywords: {Visual servoing;Cameras;Error correc- tion;Robot vision systems...

    2007

  15. [23]

    2 1/2 D visual servoing: a possible solution to improve image-based and position -based visual servoings,

    F. Chaumette and E. Malis, "2 1/2 D visual servoing: a possible solution to improve image-based and position -based visual servoings," Proceedings 2000 ICRA. Millennium Conference. IEEE International Conference on Robotics and Automation. Symposia Pro- ceedings (Cat. No.00CH37...

  16. [24]

    Fast Model Predictive Image-Based Visual Servoing for Quadrotors,

    P. Roque, E. Bin, P. Miraldo and D. V. Dimarogonas, "Fast Model Predictive Image-Based Visual Servoing for Quadrotors," 2020 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), Las Vegas, NV, USA, 2020, pp. 7566 -7572, doi: 10.1109/IROS45743.2020.934075...

    2020

  17. [25]

    Zhu, X., Bai, Y., Yu, C. et al. A new computational approach for optimal control of switched systems. J Inequal Appl 2024, 53 (2024). https://doi.org/10.1186/s13660-024-03124-2

    2024 doi

  18. [26]

    2 1/2 D Visual Servoing,

    E. Malis, F. Chaumette, and S. Boudet, "2 1/2 D Visual Servoing," 1999

    1999

  19. [27]

    Robust uncalibrated visual servoing control based on disturbance observer,

    Z. Ma and J. Su, "Robust uncalibrated visual servoing control based on disturbance observer," ISA Transactions, vol. 59, pp. 193– 204, 2015. doi: 10.1016/j.isatra.2015.07.003

    2015 doi

  20. [28]

    Modern Wiener-Hopf Design of Optimal Controllers-Part II: The Multivariable Case

    Youla, D., Jabr, H., and Bongiorno, J. Modern Wiener-Hopf Design of Optimal Controllers-Part II: The Multivariable Case. IEEE Transactions on Automatic Control.1976; 21(3):319-338. DOI: 10.1109/TAC.1976.1101223

    1976

  21. [29]

    Illingworth J., Kittler J.: A survey of the Hough transform. Comput. Vis. Graph. Image Process. 44(1), 87–116 (1988)

    1988

  22. [30]

    A kinematic notation for lower-pair mechanisms based on matri- ces

    Denavit, Jacques; Hartenberg, Richard Scheunemann (1955). A kinematic notation for lower-pair mechanisms based on matri- ces. Trans ASME J. Appl. Mech. .1955;23 (2): 215–221. DOI:10.1115/ 1.4011045

    1955

  23. [31]

    Kim, M.-Y., Cho, H., & Kim, J. H. (2001). Hand/eye calibration of a robot arm with a 3D visual sensor. Proceedings of SPIE - The International Society for Optical Engineering, 4564. https://doi.org/10.1117/12.444081

    2001 doi

  24. [32]

    Evangelista, D., Allegro, D., Terreran, M., Pretto, A., & Ghidoni, S. (2022). An unified iterative hand-eye calibration method for eye-on-base and eye-in-hand setups. In Proceedings of the 2022 IEEE 27th International Conference on Emerging Technologies and Factory Automation (...

    2022

  25. [33]

    H., & Hirzinger, G

    Strobl, K. H., & Hirzinger, G. (2006). Optimal hand -eye calibration. In Proceedings of the 2006 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) (pp. 4647–4653). IEEE. https://doi.org/10.1109/IROS.2006.282250ResearchGate

    2006

  26. [34]

    ABB IRB 4600 -40/2.55 Product Manual [Internet]

    Anonymous. ABB IRB 4600 -40/2.55 Product Manual [Internet]. 2013. Available from: https://www.manualslib.com/man- ual/1449302

    2013

  27. [35]

    Mark, W.S., M.V. (1989). Robot Dynamics and control. John Wiley & Sons, Inc. 1989

    1989

  28. [36]

    Innovative Adaptive Imaged Based Visual Servoing Control of 6DoFs Industrial Robot Ma- nipulators

    R. Li and F. Assadian Chapter 5: “Innovative Adaptive Imaged Based Visual Servoing Control of 6DoFs Industrial Robot Ma- nipulators” in Book Latest Adaptive Control System. ISBN: 9780854667123

  29. [37]

    Robust Control: Youla Parameterization Approach

    F. Assadian and K. Mallon “Robust Control: Youla Parameterization Approach” John Wiley & Sons, Ltd, 2022, ch. 10, pp. 217-

    2022

  30. [38]

    Image De -noising by Various Filters for Different Noise

    Patidar, P., Gupta, M., Srivastava, S., & Nagawat, A.K. Image De -noising by Various Filters for Different Noise. International Journal of Computer Applications. 2010; 9: 45-50

    2010

  31. [39]

    Stereolabs Docs: API Reference, Tutorials, and Integration

    Anonymous. Stereolabs Docs: API Reference, Tutorials, and Integration. Available from: https://www.stereolabs.com/docs [Ac- cessed: 2024-7-18]

    2024

  32. [40]

    Signal Averaging for Noise Reduction

    Ng J., Goldberger J.J. Signal Averaging for Noise Reduction. In: Goldberger J., Ng J. (eds) Practical Signal and Image Processing in Clinical Cardiology. London: Springer; 2010. p. 69-77. DOI: 10.1007/978-1-84882-515-4.ch7

    2010 doi

  33. [41]

    Zhao and H

    R. Zhao and H. Cui. Improved threshold denoising method based on wavelet transform. 2015 7th International Conference on Modelling, Identification and Control (ICMIC); 2015: pp. 1-4. DOI: 10.1109/ICMIC.2015.7409352

    2015

  34. [43]

    C., & Glover, K

    Zhou, K., Doyle, J. C., & Glover, K. (1996). Robust and Optimal Control. Prentice Hall. Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual au- thor(s) and contributor(s) and not of MDPI and/or the edi...

    1996

  35. [246]

    doi: 10.1002/9781119500292.ch10

    doi

Pith tools

Reviewed August 7, 2026 · model on record in the stance chip above.