EUVIMEDEUVIMED entdecken
Crossref

Intelligent adaptive MIMO predictive control using stacked polynomial fuzzy broad learning systems

Ali Rospawan, Ching-Chih Tsai

ISA Transactions · 2026-10

DOI: 10.1016/j.isatra.2026.10.010 ↗

Kurz gefasst

Fachartikel in der Zeitschrift ISA Transactions, 2026-10. Verfasst von Ali Rospawan und 1 weiteren Autor:innen. Beteiligte Einrichtung: Department of Electrical Engineering, National Chung Hsing University, Taichung 40227, Taiwan; Department of Electronics Engineering, Politeknik Manu…. Die Arbeit nennt 49 Literaturangaben.

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

This paper presents a novel adaptive nonlinear model predictive control (ANMPC) framework that integrates polynomial fuzzy neural networks (PFNN) with stacked broad learning systems (SBLS) for setpoint tracking of nonlinear coupled and time-delay MIMO systems. The proposed SPFBLS-ANMPC leverages the strengths of PFNN for state estimation and SBLS for capturing complex input-output relationships via hierarchical feature learning. The convergence and boundedness of the SPFBLS learning and the closed-loop stability of the overall control system are theoretically established through five theorems, complemented by a structure-specific corollary yielding a closed-form, real-time-implementable learning-rate bound. Comprehensive validation via simulation and real-time experiments shows significant performance improvements over conventional PID control, with up to 37.99% improvement in ISE for the two-tank flow control system and up to 52.95% improvement in ITAE for the coupled, pulsed-load-disturbed extrusion barrel system, while also outperforming model reference adaptive control (MRAC), MRAC with integral feedback (MRAC-I), fuzzy- and fuzzy-BLS-based adaptive controllers, and incremental MPC (IMPC) on all tracking metrics. Real-time deployment on a hybrid MATLAB-STM32 architecture confirms the practical feasibility and robustness under real-world operating conditions with noise, disturbances, and model uncertainties. The results establish SPFBLS-ANMPC as a promising approach for advanced nonlinear control in industrial automation.

Schlagwörter

Stichwörter der Autor:innen

Extrusion barrelModel predictive controlPolynomial fuzzy neural networkStacked broad learning systemsTemperature controlTwo-tank level system

Autor:innen und Institutionen

  • Department of Electrical Engineering, National Chung Hsing University, Taichung 40227, Taiwan; Department of Electronics Engineering, Politeknik Manufaktur Negeri Bangka Belitung, Sungailiat 33211, Indonesia. Electronic…
  • Department of Electrical Engineering, National Chung Hsing University, Taichung 40227, Taiwan. Electronic address: cctsai@nchu.edu.tw.

Zitieren

APA

Rospawan, A., & Tsai, C. C. (2026). Intelligent adaptive MIMO predictive control using stacked polynomial fuzzy broad learning systems. ISA Transactions. https://doi.org/10.1016/j.isatra.2026.10.010

Vancouver

Rospawan A, Tsai CC. Intelligent adaptive MIMO predictive control using stacked polynomial fuzzy broad learning systems. ISA Transactions. 2026. doi:10.1016/j.isatra.2026.10.010

BibTeX
@article{Rospawan2026,
  author = {Rospawan, Ali and Tsai, Ching-Chih},
  title = {Intelligent adaptive MIMO predictive control using stacked polynomial fuzzy broad learning systems},
  journal = {ISA Transactions},
  year = {2026},
  doi = {10.1016/j.isatra.2026.10.010}
}
Kostenlos starten →