Learning Control-Affine Reduced-Order Models via Autoencoders
2026-06-03 • Machine Learning
Machine Learning
AI summaryⓘ
The authors propose a method to simplify complex systems by using autoencoders to shrink the system's high-dimensional data into smaller, manageable parts that still allow for control. They train the autoencoder and a model of the system's behavior together to make accurate predictions. They also improve the model by considering past states and inputs, which helps in making better predictions while keeping the system easy to control. Their approach was tested on examples and compared to simpler models, showing how well it predicts and controls system behavior.
autoencoderreduced-order modelcontrol-affine systemstate-space dynamicsfeedback linearizationlatent spacesequence-based modelprediction accuracysystem controlmodel training
Authors
Ali Mjalled, Martin Mönnigmann
Abstract
We present in this paper a framework for the identification of control-affine reduced-order models (ROMs). The proposed method utilizes autoencoders (AEs) to transform the high-dimensional states, and potentially the high-dimensional inputs, into reduced latent ones suitable for control-affine state-space dynamics. This is achieved by simultaneous training of the AE and the state-space model. In addition, we extend the discrete ROM formulation to a sequence-based model, which processes state and input histories to improve prediction accuracy while preserving the control-affine structure. We motivate our framework by applying feedback linearization to the derived models, and we present guidelines for its efficient use. The proposed framework is assessed on two numerical examples and its performance is compared to a baseline model, where the AE identifies a latent space with linear state-space dynamics. The assessment involves evaluating the prediction accuracy of the ROM on test data and its effectiveness in controlling the system to a desired state or trajectory.