Visual feedback vibration control of flexible hinged plate system based on reinforcement learning algorithm

[Display omitted] •A fixed flexible hinged plate system is designed and FEM is obtained.•Wavelet analysis and SA optimization algorithm are used to identify the model parameters.•Laser displacement sensors and binocular vision are used to detect vibration and feedback control.•DAPO RL algorithm is d...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Veröffentlicht in:Mechanical systems and signal processing Jg. 224; S. 112005
Hauptverfasser: Qiu, Zhi-cheng, Liu, Yi-hong
Format: Journal Article
Sprache:Englisch
Veröffentlicht: Elsevier Ltd 01.02.2025
Schlagworte:
Divergence-augmented policy optimization (DAPO)
Flexible hinged plate
Model identification
Reinforcement learning algorithm
Vibration suppression
Visual detection
Divergence-augmented policy optimization (DAPO)
Flexible hinged plate
Vibration suppression
Visual detection
Model identification
Reinforcement learning algorithm
ISSN:0888-3270
Online-Zugang:Volltext
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
Beschreibung
Zusammenfassung:[Display omitted] •A fixed flexible hinged plate system is designed and FEM is obtained.•Wavelet analysis and SA optimization algorithm are used to identify the model parameters.•Laser displacement sensors and binocular vision are used to detect vibration and feedback control.•DAPO RL algorithm is designed to train the modal controllers for vibration suppression.•Simulation and experiments are carried out to verify the effectiveness of the DAPO modal controllers. In order to simulate the bending and torsional coupled vibration of solar panels under external disturbances, an experimental system of a fixed flexible hinged plate is established. The dynamic model is established by finite element method (FEM), and the model parameters are accurately identified by wavelet transform and simulated annealing (SA) optimization method. Two laser displacement sensors and binocular vision are combined to detect the bending and torsional vibrations and used for feedback control. A combined piezoelectric (PZT) control scheme is designed to suppress the bending and torsional vibrations of the flexible hinged plate. A vibration control method based on divergence augmented policy optimization (DAPO) reinforcement learning (RL) for the fixed flexible hinged plate is developed to train the modal controllers. Simulation and experiments are carried out to verify the effectiveness of the applied vibration control scheme and algorithm. The simulation and experimental results show that DAPO RL modal controllers have better vibration suppression effect than large gain proportional derivative (PD) controllers, especially for small amplitude residual vibration.
ISSN:0888-3270
DOI:10.1016/j.ymssp.2024.112005