Generalized performance of neural network controllers for feedforward active control of nonlinear systems
Generalized performance of neural network controllers for feedforward active control of nonlinear systems
Over the past few decades, advances in digital technologies have allowed for the development of complex active control solutions for both vibration and acoustic control and have been utilised in a wide range of applications. Such control systems are commonly designed using linear filters which cannot fully capture the dynamics of nonlinear systems. To overcome such issues, it has previously been shown that replacing linear controllers with Neural Networks (NNs) can improve control performance in the presence of nonlinearities in both the system plant and primary path. However, the performance of the controller across excitation levels has not been frequently explored. Controllers with good performance across a range of excitation levels would be essential in the control of many real systems. In this paper, a method of training Multilayer Perceptrons (MLPs) for single-input-single-output (SISO) feedforward acoustic noise control is presented. In a simple time-discrete simulation, the performance of the trained NNs is investigated for different excitation levels. The effects of the properties of the training data and NN controller on generalized performance are explored.
machine learning, active control, neural network, nonlinear
Pike, Alexander
1cd3f629-7971-4b9c-9b4a-636df608bbe0
Cheer, Jordan
8e452f50-4c7d-4d4e-913a-34015e99b9dc
14 September 2023
Pike, Alexander
1cd3f629-7971-4b9c-9b4a-636df608bbe0
Cheer, Jordan
8e452f50-4c7d-4d4e-913a-34015e99b9dc
Pike, Alexander and Cheer, Jordan
(2023)
Generalized performance of neural network controllers for feedforward active control of nonlinear systems.
10th Convention of the European Acoustics Association: Forum Acusticum 2023: acoustics for a green world, Politecnico di Torino, Torino, Italy.
11 - 15 Sep 2023.
8 pp
.
Record type:
Conference or Workshop Item
(Paper)
Abstract
Over the past few decades, advances in digital technologies have allowed for the development of complex active control solutions for both vibration and acoustic control and have been utilised in a wide range of applications. Such control systems are commonly designed using linear filters which cannot fully capture the dynamics of nonlinear systems. To overcome such issues, it has previously been shown that replacing linear controllers with Neural Networks (NNs) can improve control performance in the presence of nonlinearities in both the system plant and primary path. However, the performance of the controller across excitation levels has not been frequently explored. Controllers with good performance across a range of excitation levels would be essential in the control of many real systems. In this paper, a method of training Multilayer Perceptrons (MLPs) for single-input-single-output (SISO) feedforward acoustic noise control is presented. In a simple time-discrete simulation, the performance of the trained NNs is investigated for different excitation levels. The effects of the properties of the training data and NN controller on generalized performance are explored.
Text
FA23_FINAL
- Version of Record
More information
Published date: 14 September 2023
Venue - Dates:
10th Convention of the European Acoustics Association: Forum Acusticum 2023: acoustics for a green world, Politecnico di Torino, Torino, Italy, 2023-09-11 - 2023-09-15
Keywords:
machine learning, active control, neural network, nonlinear
Identifiers
Local EPrints ID: 483122
URI: http://eprints.soton.ac.uk/id/eprint/483122
PURE UUID: 2db4ee35-6aa8-4542-8436-55d4c57ec05e
Catalogue record
Date deposited: 25 Oct 2023 02:16
Last modified: 18 Mar 2024 03:17
Export record
Download statistics
Downloads from ePrints over the past year. Other digital versions may also be available to download e.g. from the publisher's website.
View more statistics
