Continuous-time system identification with neural networks: Model structures and fitting criteria

TL;DR

This paper introduces neural network-based continuous-time system identification methods using specialized model structures and fitting criteria, demonstrated through multiple case studies including benchmark datasets.

Contribution

It proposes a novel framework combining continuous-time state-space models with neural networks and custom fitting criteria for improved dynamical system learning.

Findings

Effective in modeling dynamical systems with neural networks

Validated on public benchmark datasets

Shows improved consistency between states and system dynamics

Abstract

This paper presents tailor-made neural model structures and two custom fitting criteria for learning dynamical systems. The proposed framework is based on a representation of the system behavior in terms of continuous-time state-space models. The sequence of hidden states is optimized along with the neural network parameters in order to minimize the difference between measured and estimated outputs, and at the same time to guarantee that the optimized state sequence is consistent with the estimated system dynamics. The effectiveness of the approach is demonstrated through three case studies, including two public system identification benchmarks based on experimental data.