Search papers, labs, and topics across Lattice.
This paper introduces a novel batch-wise learning pipeline and a direct recursive identification algorithm for online adaptation of neural state-space models (ANN-SS). By leveraging a subspace encoder, the proposed method achieves high accuracy while maintaining computational efficiency, addressing a significant gap in the online learning of these models. Extensive simulations validate the convergence and performance of the approach, showcasing its potential for real-time system identification tasks.
Online adaptation of neural state-space models can now be achieved with high accuracy and computational efficiency, filling a critical gap in real-time system identification.
Recent advances in deep-learning-based nonlinear system identification have led to encoder-based estimation of neural state-space (ANN-SS) models that achieve state-of-the-art performance in offline settings by estimating initial model states from past input-output data. These methods are typically used in multiple-shooting-based offline identification, and online learning of these models remains largely unexplored. This paper presents a batch-wise learning pipeline and a direct recursive identification algorithm for subspace encoder-based ANN-SS models. We provide convergence analysis of the recursive formulation and validate its performance through extensive simulation studies. The results demonstrate that the proposed approach enables computationally efficient online adaptation with high model accuracy.
