Inference and Learning of Nonlinear LFR State-Space Models
Abstract
Estimating the parameters of nonlinear block-oriented state-space models from input-output data typically involves solving a highly non-convex optimization problem, which is prone to poor local minima and slow convergence. This paper presents a computationally efficient initialization method for nonlinear linear fractional representation (NL-LFR) models using periodic data. By first inferring the latent signals and subsequently estimating the model parameters, the approach generates initial estimates for use in a later nonlinear optimization step. The proposed method shows robustness against poor local minima, and achieves a twofold error reduction compared to the state-of-the-art on a challenging benchmark dataset.
Cite
@article{arxiv.2503.14409,
title = {Inference and Learning of Nonlinear LFR State-Space Models},
author = {Merijn Floren and Jean-Philippe Noël and Jan Swevers},
journal= {arXiv preprint arXiv:2503.14409},
year = {2025}
}
Comments
Code is available at: https://github.com/merijnfloren/freq-statespace ; final, published paper in IEEE Xplore: https://ieeexplore.ieee.org/abstract/document/11037476/