This paper proposes a sequential three-step data-driven framework to design a complete controller-allocator scheme directly from noisy input-state data, bypassing parametric system identification. First, a gain-scheduled state-feedback controller is designed to stabilize the Linear Parameter-Varying (LPV) plant. Second, a constant basis for the input matrix kernel is extracted directly from the dataset, allowing the allocator structure for output invisibility, meaning that control effort is redistributed among actuators without altering the system output. Third, a parameter-independent dynamic allocator gain is designed to regulate the allocation convergence speed, providing a robust design across the entire scheduling space. The three design steps are entirely formulated through data-based Linear Matrix Inequalities (LMIs). This combined approach guarantees closed-loop quadratic stability and minimizes an upper bound on the total actuator energy consumption over time. The effectiveness of the framework is validated through a numerical example on a randomly generated strongly input redundant LPV system, demonstrating a clear reduction in control energy while preserving closed-loop stability.
Sequential Data-Driven Design of a Complete Controller-Allocator Scheme for Strongly Input Redundant LPV Systems / Faleschini, M., Cristofaro, A.. - (2026), pp. 878-883. (34rd Mediterranean Conference on Control and Automation, MED 2026 Ancona; Italy ) [10.1109/MED70602.2026.11598316].
Sequential Data-Driven Design of a Complete Controller-Allocator Scheme for Strongly Input Redundant LPV Systems
Michelangelo Faleschini
Primo
;
2026
Abstract
This paper proposes a sequential three-step data-driven framework to design a complete controller-allocator scheme directly from noisy input-state data, bypassing parametric system identification. First, a gain-scheduled state-feedback controller is designed to stabilize the Linear Parameter-Varying (LPV) plant. Second, a constant basis for the input matrix kernel is extracted directly from the dataset, allowing the allocator structure for output invisibility, meaning that control effort is redistributed among actuators without altering the system output. Third, a parameter-independent dynamic allocator gain is designed to regulate the allocation convergence speed, providing a robust design across the entire scheduling space. The three design steps are entirely formulated through data-based Linear Matrix Inequalities (LMIs). This combined approach guarantees closed-loop quadratic stability and minimizes an upper bound on the total actuator energy consumption over time. The effectiveness of the framework is validated through a numerical example on a randomly generated strongly input redundant LPV system, demonstrating a clear reduction in control energy while preserving closed-loop stability.| File | Dimensione | Formato | |
|---|---|---|---|
|
Falaschini_Sequential-data-driven_2026.pdf
solo gestori archivio
Tipologia:
Versione editoriale (versione pubblicata con il layout dell'editore)
Licenza:
Tutti i diritti riservati (All rights reserved)
Dimensione
713.98 kB
Formato
Adobe PDF
|
713.98 kB | Adobe PDF | Contatta l'autore |
I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.


