CoChat

Data-driven control of hydraulic servo actuator: An event-triggered adaptive dynamic programming approach

Open access
Djordjevic, Vladimir, Tao, Hongfeng, Song, Xiaona, He, Shuping, Gao, Weinan, Stojanovic, Vladimir · 2023 · Mathematical Biosciences and Engineering

The DOI matches its Crossref record; 50 of the 50 references it cites were found in a registry or an index; 1 of the 45 sources its claims cite were checked for support.

Abstract

This paper considers an event-triggered learning control problem of the HSA with unknown dynamics based on adaptive dynamic programming (ADP) via output feedback.

Overview

CoChat Summary

TL;DR

This paper develops an event-triggered adaptive dynamic programming (ADP) controller for a hydraulic servo actuator (HSA) with unknown dynamics, using only measured input/output data and output feedback. Simulations show convergence of the learned policy and about 54% reduced communication versus periodic ADP control.

Key findings

  • The proposed ADP-based online learning controller avoids requiring knowledge of the HSA dynamics or exosystem dynamics, using state reconstruction from measured input/output data instead of unmeasurable states and unknown parameters.
  • The event-triggered mechanism updates the control input only when the squared norm of the triggering error reaches a threshold, and the simulation reports about 54% communication reduction between the controller and the HSA compared with the ADP method using periodic sampling.
  • In simulation, the approximated cost function \(V_7\) after seven iterations is remarkably reduced relative to the initial cost \(V_1\), and the approximation errors for the cost function and optimal input are close to zero.
  • Similar control effects are achieved by the event-triggered ADP method and the periodic ADP method, while the event-triggered method keeps the control input constant between updates.

Methods

  • The HSA is modeled first as a nonlinear continuous-time system, then linearized around an operating point and discretized with a zero-order hold using sampling period \(h = 0.1\) s; the system matrices \(A, B, C\) are assumed unknown for control design but controllable and observable.
  • The controller uses ADP via output feedback: state reconstruction from input/output sequences over a horizon, critic/policy evaluation to solve Lyapunov equations, and actor/policy improvement to update the control gain; an event-triggering condition governs control updates.
  • Simulation parameters include viscous friction \(B_C = 200\) N s m\(^{-1}\), supply pressure \(p_S = 45\) bar, tank pressure \(p_0 = 1.6\) bar, bulk modulus \(e = 2 \times 10^8\) Pa, total mass \(m = 25\) kg, initial chamber volumes \(V_{a0}=V_{b0}=8.2 \times 10^{-6}\) m\(^3\), load spring gradient \(K_e = 10^1\), effective areas \(A_a = 4.91 \times 10^{-4}\) m\(^2\) and \(A_b = 2.43 \times 10^{-4}\) m\(^2\), internal leakage coefficient \(c_{Li}=5 \times 10^{-14}\), piston stroke \(L=1\) m, and discharge coefficients \(c_{vi}=1.15\); \(Q\) and \(R\) are identity matrices, observability index \(N=3\), initial state \(x_0 = [5\ 5\ 10]\), and convergence threshold \(\epsilon = 10^{-1}\).

Limitations

  • The paper does not state an explicit limitations section.
  • Evident from the text: validation is simulation-only; no experimental implementation on a physical HSA is reported.
  • Evident from the text: although the controller design does not require exact HSA matrices, the numerical verification assumes the system matrices in (2.10) and (2.11) are known.
  • Evident from the text: the method relies on a linearized/discrete-time HSA model and on exploration noise satisfying persistent excitation, and the paper notes that selecting exploration noise is non-trivial and can affect solution accuracy.

Model-written summary. Check it against the paper before relying on it.

Integrity signals

Registry

Crossref · 10.3934/mbe.2023376

Cites

50 of 50 references verified0 citation errors · 0 not checkable

Statistics

Nothing to recompute.

Figures

Figure 1 The HSA configuration.
Figure 1 The HSA configuration.From the PDF
Figure 2 Event-triggered ADP-based control algorithm for the discretized HSA system.
Figure 2 Event-triggered ADP-based control algorithm for the discretized HSA system.From the PDF
Figure 3 Flowchart of ADP-based controller design.
Figure 3 Flowchart of ADP-based controller design.From the PDF
Figure 4 Hybrid nature of control signal.
Figure 4 Hybrid nature of control signal.From the PDF
Figure 5 Convergence of ¯ Pj and ¯ Kj to their respective optimal values ¯ P and ¯ K during the learning process.
Figure 5 Convergence of ¯ Pj and ¯ Kj to their respective optimal values ¯ P and ¯ K during the learning process.From the PDF
Figure 6 (a) Comparison of the cost functions during learning; (b) error between the optimal and approximated cost function signal.
Figure 6 (a) Comparison of the cost functions during learning; (b) error between the optimal and approximated cost function signal.From the PDF
Figure 7 (a) Comparison of the control policies during the learning process; (b) error between the optimal and approximated input signal.
Figure 7 (a) Comparison of the control policies during the learning process; (b) error between the optimal and approximated input signal.From the PDF
Figure 8 Control input and states of the HSA model by using the ADP-based controller.
Figure 8 Control input and states of the HSA model by using the ADP-based controller.From the PDF
Figure 9 Control input and states of the HSA model by using the event-triggered ADPbased controller.
Figure 9 Control input and states of the HSA model by using the event-triggered ADPbased controller.From the PDF
Figure 10 Comparison of the total sampling numbers.
Figure 10 Comparison of the total sampling numbers.From the PDF
Figure 11 Sequence of steps of event-triggered sampling.
Figure 11 Sequence of steps of event-triggered sampling.From the PDF

Verification

Claims the paper makes

1 Valid · 0 Not valid · 22 Not checked · 28 Own results and interpretations

28 more claims

Create a free account to see all 51 claims and to verify the 2 references still open.

See all claims — free account
  • It may state two separate claims as one. 12
  • Probably not a substantial claim. 9
  • The quote only partly states this claim. 1

Create a free account to see these claims.

Citations

Cites

Research works printed but not linked: 7

  1. N. Manring, Fluid Power Pumps and Motors: Analysis, Design, and Control, 1$^{st}$ edition, McGraw-Hill Education, New York, 2013.
  2. D. Bertsekas, Reinforcement Learning and Optimal Control, 1$^{st}$ edition, Athena Scientific, Belmont, Massachusetts, 2019.
  3. D. Bertsekas, Dynamic Programming and Optimal Control: Volume I, 4$^{th}$ edition, Athena Scientific, Belmont, Massachusetts, 2012.
  4. P. J. Werbos, Beyond Regression: New Tools for Prediction and Analysis in the Behavioral Sciences, Ph.D thesis, Harvard University, 1974.
  5. L. Ljung, System Identification: Theory for the User, 2$^{nd}$ edition, Prentice-Hall, Upper Saddle River, New Jersey, 1999.
  6. K. J. Åström, B. Wittenmark, Adaptive Control, 2$^{nd}$ edition, Dover Publication Inc, Mineola, New York, 2008.
  7. P. A. Ioannou, J. Sun, Robust Adaptive Control, 2$^{nd}$ edition, Dover Publication Inc, Mineola, New York, 2012.

Cited by

See all 107

Metadata

Authors
Djordjevic, V., Tao, H., Song, X., He, S., Gao, W., & Stojanovic, V.
Affiliations
  • University of Kragujevac (RS)
  • Jiangnan University (CN)
  • Henan University of Science and Technology (CN)
  • Anhui University (CN)
  • State Key Laboratory of Synthetical Automation for Process Industries (CN)
  • Faculty of Mechanical and Civil Engineering, University of Kragujevac, 36000 Kraljevo, Serbia
Venue
Mathematical Biosciences and Engineering
Published
2023
Volume
20
Issue
5
Pages
8561-8582
ISSN
  • 1547-1063 (ISSN-L)
  • 1551-0018
OpenAlex
W4323338927
Cited by
107
References
44
Citations by year
2026: 7 · 2025: 15 · 2024: 51 · 2023: 32
FWCI
15.03
Type
journal article
Language
en
Open access
Licence
cc-by-nc-sa
Locations
  • Mathematical Biosciences & Engineering (publishedVersion, cc-by-nc-sa)
  • PubMed (publishedVersion)
  • DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion)
Funders
  • National Natural Science Foundation of China (B23008, 2022i01020013, 61773181, 62203153, 62103293, JUSRP51733B, 61976081, 62073001, 202301)
  • Ministarstvo Prosvete, Nauke i Tehnološkog Razvoja (451-03-47/2023-01/200108, 451-03-47/, 451-03-47/2023-01)
  • Higher Education Discipline Innovation Project (JUSRP51733B, B23008)
  • Fundamental Research Funds for the Central Universities (B23008, JUSRP51733B)
Topics
  • Physical Sciences › Engineering › Control and Systems Engineering › Adaptive Dynamic Programming Control
  • Physical Sciences › Engineering › Mechanical Engineering › Hydraulic and Pneumatic Systems
  • Physical Sciences › Engineering › Control and Systems Engineering › Adaptive Control of Nonlinear Systems
Keywords
hydraulic servo actuator, event-triggered control, adaptive dynamic programming, data-driven control, output feedback, unknown dynamics, nonlinear systems, optimal control, convergence analysis, communication resource reduction

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

The claim verdicts this report counts are experimental: their accuracy is not yet measured.

A language model drew the claims from the paper on 2026-10-10, under version 3 of the claim definition (model deepseek-v4.1-flash, prompt extract-v3).

A second model labelled each claim on 2026-10-10 (model jev-1.13-20260917, definition revision gold-guideline-v0.2).

Cited sources are checked with a language model’s help (qwen3.8-omni-flash).

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review

Data sources

Bibliographic data: Semantic Scholar, OpenAlex, Crossref.

Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).

Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.