Physics-Informed Learning of Feedback-Linearizing Representations
My rating
Feedback linearization relies on finding a linearizing coordinate transform, but the governing PDEs become computationally intractable for large system dimensions. This work proposes a cascaded physics-informed neural network (PINN) framework to approximate the solutions of these PDEs. The resource targets physics-informed machine learning for nonlinear control systems rather than a specific physical flow field, bridging PDE-based control theory and neural-network-based scientific computation.
Tags: