pith. sign in

arxiv: 2604.13763 · v1 · submitted 2026-04-15 · 📡 eess.SY · cs.AI· cs.LG· cs.NE· cs.RO· cs.SY

A Dynamic-Growing Fuzzy-Neuro Controller, Application to a 3PSP Parallel Robot

Mohsen Jalaeian-Farimani , Mohammad-R Akbarzadeh-T , Alireza Akbarzadeh , Mostafa Ghaemi This is my paper

Pith reviewed 2026-05-10 12:54 UTC · model grok-4.3

classification 📡 eess.SY cs.AIcs.LGcs.NEcs.ROcs.SY
keywords dynamic growing fuzzy neural controller3PSP parallel robotadaptive controlsliding mode controlposition controlself-organizing fuzzy systemsrobot controlparameter variation
0
0 comments X

The pith

A dynamic-growing fuzzy-neural controller with adaptation controls the 3PSP robot faster using less computation while ensuring stability.

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

The paper develops a Dynamic Growing Fuzzy Neural Controller (DGFNC) for position control of a 3PSP parallel robot, which features complex dynamics. It grows the rule base conservatively without a pruning step and instead uses an adaptive strategy to cope with parameter variations, backed by a sliding mode controller for stability. This combination is intended to deliver quicker responses at lower computational cost than traditional self-organizing fuzzy-neuro systems. Readers interested in industrial robotics would care because such controllers could simplify real-time implementation in systems where dynamics change over time.

Core claim

The authors claim that their DGFNC, by adding new rules more conservatively than other methods and omitting pruning in favor of adaptation, combined with a sliding mode nonlinear controller, achieves faster response with less computation while maintaining overall stability when applied to the 3PSP parallel robot, as supported by several simulations.

What carries the argument

The Dynamic Growing Fuzzy Neural Controller (DGFNC) with its conservative rule-adding mechanism, paired with an adaptive strategy to handle variations and a sliding mode controller for guaranteed stability.

If this is right

  • The strategy maintains system stability through the sliding mode-based nonlinear controller.
  • It provides faster response times compared to other self-organizing approaches.
  • Computational load is reduced by avoiding the need for a pruning mechanism.
  • The approach is suitable for industrial systems with complex and varying dynamics like the 3PSP robot.
  • Simulations confirm the merits of the DGFNC strategy for position control.

Where Pith is reading between the lines

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

  • Similar conservative growing mechanisms could benefit other adaptive controllers in robotics to prevent unnecessary complexity.
  • Testing on hardware beyond simulations would be needed to confirm real-world gains in speed and computation.
  • The method might extend to other parallel manipulators or systems requiring robust position control under uncertainty.
  • Without pruning, the rule base could grow larger over time, potentially affecting long-term efficiency in highly variable conditions.

Load-bearing premise

The adaptive strategy without pruning suffices to manage all parameter variations in the robot's dynamics, and the simulation results translate directly to real-world performance.

What would settle it

A real-world test or simulation with unaccounted parameter variations where the controller loses stability or requires more computation than claimed would disprove the central claim.

Figures

Figures reproduced from arXiv: 2604.13763 by Alireza Akbarzadeh, Mohammad-R Akbarzadeh-T, Mohsen Jalaeian-Farimani, Mostafa Ghaemi.

Figure 1
Figure 1. Figure 1: DGFNC close-loop control block diagram. Authorized licensed use limited to: Politecnico di Milano. Downloaded on March 05,2026 at 10:48:49 UTC from IEEE Xplore. Restrictions apply [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: Virtual structure of the special 3-PSP parallel manipulator (This figure is illustrated by Solid Works software based on 3PSP robots model). The inverse Kinematics objective is obtaining the position of the actuators for a given pose of the end-effector (moving star shape). The control’s objective is computing the torque of motors for its determined positions which is obtained by inverse kinematic equation… view at source ↗
Figure 6
Figure 6. Figure 6: Zoomed desired and actual motors angle [PITH_FULL_IMAGE:figures/full_fig_p005_6.png] view at source ↗
Figure 7
Figure 7. Figure 7: Tool desired and actual trajectories (Helix) [PITH_FULL_IMAGE:figures/full_fig_p005_7.png] view at source ↗
Figure 5
Figure 5. Figure 5: Desired and actual motors angle (Circle trajectory) [PITH_FULL_IMAGE:figures/full_fig_p005_5.png] view at source ↗
read the original abstract

To date, various paradigms of soft-Computing have been used to solve many modern problems. Among them, a self organizing combination of fuzzy systems and neural networks can make a powerful decision making system. Here, a Dynamic Growing Fuzzy Neural Controller (DGFNC) is combined with an adaptive strategy and applied to a 3PSP parallel robot position control problem. Specifically, the dynamic growing mechanism is considered in more detail. In contrast to other self-organizing methods, DGFNC adds new rules more conservatively; hence the pruning mechanism is omitted. Instead, the adaptive strategy 'adapts' the control system to parameter variation. Furthermore, a sliding mode-based nonlinear controller ensures system stability. The resulting general control strategy aims to achieve faster response with less computation while maintaining overall stability. Finally, the 3PSP is chosen due to its complex dynamics and the utility of such approaches in modern industrial systems. Several simulations support the merits of the proposed DGFNC strategy as applied to the 3PSP robot.

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

3 major / 1 minor

Summary. The manuscript proposes a Dynamic-Growing Fuzzy-Neuro Controller (DGFNC) that incorporates an adaptive strategy in place of pruning and is augmented by a sliding-mode term, applied to position control of the 3PSP parallel robot. The central claim is that the conservative rule-growing mechanism plus adaptation yields faster response, lower computational cost, and maintained stability for the robot's complex nonlinear dynamics, with these merits supported by several simulations.

Significance. If the performance advantages can be substantiated with quantitative metrics and bounded rule growth, the work would offer a practical neuro-fuzzy control variant for parallel robots that reduces overhead relative to standard self-organizing approaches while retaining Lyapunov-based stability guarantees.

major comments (3)
  1. [Simulation results] Simulation results: The abstract states that 'several simulations support the merits' of faster response and less computation, yet no quantitative metrics (settling time, overshoot, execution time, or FLOPs), rule-growth statistics over time, disturbance amplitudes, or comparisons against pruned baselines or other self-organizing fuzzy-neuro controllers are supplied. This leaves the central performance claims unverifiable.
  2. [Controller design] Adaptive strategy section: The decision to omit pruning is justified by the claim that the adaptive strategy alone 'adapts' the system to parameter variations, but no explicit bound on rule cardinality, sensitivity analysis on adaptation gains, or demonstration that rule growth remains controlled under sustained parameter drifts in the 3PSP's coupled dynamics is provided, undermining the 'less computation' assertion.
  3. [Stability analysis] Stability analysis: The sliding-mode term is asserted to ensure overall stability, but the manuscript does not detail the interaction between dynamically added fuzzy-neuro rules (without pruning) and the sliding-mode surface, nor does it supply a composite Lyapunov function or robustness margins that account for rule growth.
minor comments (1)
  1. [Abstract] The abstract and introduction repeat the performance goals without defining the quantitative criteria (e.g., response time reduction percentage) used to evaluate them.

Simulated Author's Rebuttal

3 responses · 0 unresolved

We thank the referee for the detailed and constructive comments on our manuscript. We have carefully addressed each major point by enhancing the simulation results with quantitative metrics, providing bounds and analysis for the adaptive strategy, and elaborating on the stability proof. These revisions strengthen the paper's claims regarding the DGFNC's performance advantages.

read point-by-point responses

  1. Referee: [Simulation results] Simulation results: The abstract states that 'several simulations support the merits' of faster response and less computation, yet no quantitative metrics (settling time, overshoot, execution time, or FLOPs), rule-growth statistics over time, disturbance amplitudes, or comparisons against pruned baselines or other self-organizing fuzzy-neuro controllers are supplied. This leaves the central performance claims unverifiable.

    Authors: We acknowledge that the original manuscript lacked explicit quantitative metrics in the abstract and main text. In the revised version, we have added a new table comparing settling times, overshoot percentages, and computational execution times (in milliseconds per control cycle) for the proposed DGFNC against standard fuzzy-neuro controllers and pruned variants. Additionally, a figure illustrates rule growth over simulation time under varying disturbances, with amplitudes specified in the text. These additions make the performance claims verifiable and demonstrate the advantages in response speed and reduced computation. revision: yes

  2. Referee: [Controller design] Adaptive strategy section: The decision to omit pruning is justified by the claim that the adaptive strategy alone 'adapts' the system to parameter variations, but no explicit bound on rule cardinality, sensitivity analysis on adaptation gains, or demonstration that rule growth remains controlled under sustained parameter drifts in the 3PSP's coupled dynamics is provided, undermining the 'less computation' assertion.

    Authors: The conservative growing mechanism in DGFNC is designed to limit rule addition based on error thresholds and novelty criteria, which inherently bounds the rule set. In the revision, we have included a mathematical bound on maximum rule cardinality derived from the adaptation laws and the 3PSP dynamics. We also provide sensitivity analysis plots for adaptation gains and a simulation demonstrating bounded growth under sustained parameter drifts, supporting the reduced computational overhead. revision: yes

  3. Referee: [Stability analysis] Stability analysis: The sliding-mode term is asserted to ensure overall stability, but the manuscript does not detail the interaction between dynamically added fuzzy-neuro rules (without pruning) and the sliding-mode surface, nor does it supply a composite Lyapunov function or robustness margins that account for rule growth.

    Authors: We appreciate this observation. The original stability analysis used a Lyapunov function for the sliding-mode controller, but the interaction with growing rules was not fully detailed. In the revised manuscript, we present a composite Lyapunov function that incorporates both the sliding-mode surface and the fuzzy-neuro approximation errors, accounting for rule additions. We also derive robustness margins showing that the conservative growth ensures the approximation error remains within bounds that preserve stability. revision: yes

Circularity Check

0 steps flagged

No circularity in derivation chain; standard techniques applied without self-referential reduction.

full rationale

The paper describes combining a dynamic-growing fuzzy-neural controller with an adaptive strategy and sliding-mode term for 3PSP robot position control. The abstract states that the growing mechanism is conservative (omitting pruning) and that adaptation handles parameter variation while sliding-mode ensures stability. No equations, derivations, or first-principles results are presented that reduce any claimed performance metric (faster response, less computation) to fitted inputs or self-definitions by construction. No self-citations are invoked as load-bearing uniqueness theorems, no ansatz is smuggled, and no predictions are statistically forced from subsets of the same data. The central claims rest on simulation support rather than tautological mappings, making the derivation self-contained against external benchmarks.

Axiom & Free-Parameter Ledger

0 free parameters · 2 axioms · 0 invented entities

Claims rest on standard domain assumptions from fuzzy control and sliding-mode theory plus the paper-specific proposal that conservative rule growth plus adaptation suffices without pruning; no free parameters or new entities are explicitly introduced in the abstract.

axioms (2)
  • domain assumption Sliding mode control ensures stability of the combined fuzzy-neural adaptive system
    Invoked in abstract to guarantee overall stability without detailed proof or conditions provided.
  • ad hoc to paper Adaptive strategy can replace pruning for handling parameter variations
    Proposed as the key difference from other self-organizing methods; no derivation or bounds given.

pith-pipeline@v0.9.0 · 5512 in / 1409 out tokens · 62011 ms · 2026-05-10T12:54:27.251799+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

24 extracted references · 24 canonical work pages

  1. [1]

    Shuanghe Yu, Xinghuo Yu, Zhihong Man, A Fuzzy Neural Network Approximator with Fast Terminal Sliding Mode and Its Application, 9 th International Conference on Neural Information Processing (ICONIP’02), 2002

    work page 2002

  2. [2]

    On the approximate realization of continuous mapping by neural networks,

    K. Funahashi, "On the approximate realization of continuous mapping by neural networks," Neural Networks, vol. 2, pp. 183-192, 1989

    work page 1989

  3. [3]

    2 (1999) 359–366

    K.Hornik, M.Stinchcombe, H.White, Multilayer feedforward networks are universal approximators, Neural Netw. 2 (1999) 359–366

    work page 1999

  4. [4]

    Analysis and design of fuzzy logic controller,

    L. X. Wang and J.M.Mendel, “Analysis and design of fuzzy logic controller,” USC SIPI Rep, 184, 1991

    work page 1991

  5. [5]

    Back propagation fuzzy systems as nonlinear dynamic system identifiers,

    L. X. Wang and J.M.Mendel, “Back propagation fuzzy systems as nonlinear dynamic system identifiers,” in Proc, IEEE 1992 Int. Conf. Fuzzy Systems (San Diego, CA) Mar 1992, pp, 1409-1418

    work page 1992

  6. [6]

    Honggui Han, Junfei Qiao, A Self-Organizing Fuzzy Neural Network Based on a Growing-and-Pruning Algorithm, IEEE TRANSACTIONS ON FUZZY SYSTEMS, VOL. 18, NO. 6, DECEMBER 2010

    work page 2010

  7. [7]

    Chun-Fei Hsu, Self-Organizing Adaptive Fuzzy Neural Control for a Class of Nonlinear Systems, IEEE TRANSACTIONS ON NEURAL NETWORKS, VOL. 18, NO. 4, JULY 2007

    work page 2007

  8. [8]

    Shiqian Wu, Meng Joo Er, Dynamic Fuzzy Neural Networks—A Novel Approach to Function Approximation, IEEE TRANSACTIONS ON SYSTEMS, MAN, AND CYBERNETICS—PART B: CYBERNETICS, VOL. 30, NO. 2, APRIL 2000

    work page 2000

  9. [9]

    Jeich Mar, Feng-Jie Lin, An ANFIS Controller for the Car-Following Collision Prevention System, IEEE TRANSACTIONS ON VEHICULAR TECHNOLOGY, VOL. 50, NO. 4, JULY 2001

    work page 2001

  10. [10]

    Neural Netw

    L.X.Wang, J.M.Mendel, Fuzzy basis functions, universal approximation, and orthogonal least squares learning, IEEE Trans. Neural Netw. 3 (1992) 807–814

    work page 1992

  11. [11]

    F.Da, W.Song, Fuzzy neural networks for direct adaptive control, IEEE Trans. Ind. Electron. 50 (2003) 507–513

    work page 2003

  12. [12]

    Auto-structuring fuzzy neural system for intelligent control

    Kuo-Hsiang Cheng, “Auto-structuring fuzzy neural system for intelligent control” Journal of the Franklin Institute 346 (2009) 267–288

    work page 2009

  13. [13]

    Wang, Adaptive Fuzzy Systems and Control: Design and Stability Analysis, Prentice-Hall, Englewood Cliffs, NJ, 1994

    L.X. Wang, Adaptive Fuzzy Systems and Control: Design and Stability Analysis, Prentice-Hall, Englewood Cliffs, NJ, 1994

    work page 1994

  14. [14]

    J . H . P a r k , S . H . H u h , S . H . K i m , S . J . S e o , G . T . P a r k , D i r e c t a d a p t i v e controller for nonaffine nonlinear systems using self-structuring neural networks, IEEE Trans. Neural Netw. 16 (2005) 414–422

    work page 2005

  15. [15]

    Neural Netw

    C.-M.Lin,L.-Y.Chen,C.-H.Chen, RCMAC hybrid control for MIMO uncertain nonlinear systems using sliding-mode technology, IEEE Trans. Neural Netw. 18 (2007) 708–720

    work page 2007

  16. [16]

    V. I. Utkin, Sliding Modes in Control and Optimization. New York: Springer-Verlag, 1992

    work page 1992

  17. [17]

    Variable structure control: A survey,

    J. Y. Hung, W. Gao, and J. C. Hung, “Variable structure control: A survey,” IEEE Trans. Ind. Electron., vol. 40, no. 1, pp. 2–22, Feb. 1993

    work page 1993

  18. [18]

    H . F . H o a , Y . K . W o ng a , A . B. Ra d, Ro b us t f uz z y tr a c ki ng c o ntr o l f o r robotic manipulators, Simulation Modelling Practice and Theory 15 (2007) 801–816

    work page 2007

  19. [19]

    Yang Zhiyong, Wu Jiang, Mei Jiangping, Motor-mechanism dynamic model based neural network optimized computed torque control of a high speed parallel manipulator, Mechatronics 17,381–390, 2007

    work page 2007

  20. [20]

    Lia, X.G

    W. Lia, X.G. Changa, F.M. Wahlb, Jay Farrellc, Tracking control of a manipulator under uncertainty by FUZZY P+ID controller, Fuzzy Sets and Systems 122 (2001) 125–137

    work page 2001

  21. [21]

    Parallel Robots

    J. P. Merlet, “Parallel Robots”, 2nd ed., Kluwer Academic Publishers, London, 2000

    work page 2000

  22. [22]

    Hui Cheng, Yiu-Kuen Yiu, Zexiang Li, Dynamics and Control of Redundantly Actuated Parallel Manipulators, IEEE/ASME TRANSACTIONS ON MECHATRONICS, VOL. 8, NO. 4, DECEMBER 2003

    work page 2003

  23. [23]

    Mohsen Jalaeian-F., Mohammad-R. Akbarzadeh-T., Ali Akbarzadeh Tootoo nchi, Damoon Azarpazhooh, 'Dynamic-Growing Fuzzy-Neural Controller, as Control System of Special 3PSP Parallel Robot', IEEE indexed, The 2nd International Conference on Control, Instrumentation, and Automation (ICCIA), 2011

    work page 2011

  24. [24]

    Close form Solutions for Inverse and Direct Position Analysis of a Special 3-PSP Parallel Manipulator,

    P. Mahmoodi nia and A. R. Akbarzadeh Tootoonchi, "Close form Solutions for Inverse and Direct Position Analysis of a Special 3-PSP Parallel Manipulator," ISME conference, 2009. Authorized licensed use limited to: Politecnico di Milano. Downloaded on March 05,2026 at 10:48:49 UTC from IEEE Xplore. Restrictions apply

    work page 2009