Estimation of Vehicle States Using a Cascaded Hybrid Estimation Method

Marvin Glomsda, Hendrik Tino Prümer, Philipp Maximilian Sieberg

2025

Abstract

Three models using a cascaded hybrid estimation method with physical models of different degrees of accuracy are evaluated for their overall precision and interpretability. Hybrid estimation methods hereby denote methods concatenating the properties of physics-based models and artificial neural networks for the purpose of improved state estimation. Cascaded hybrid estimation methods are a subtype of these methods, combining a physical model and an artificial neural network in a way that one acts as the input of the other. In this publication the result of a physical model is fed into a neural network to improve the estimation quality. It can be shown that the degree of accuracy of the physical model has an influence on the overall estimation quality, with more accurate physical models yielding better results, but less accurate models can provide a more significant improvement through the artificial neural network. This is likely due to the larger residual error that can be used to train the artificial neural network.

Paper Citation

in Harvard Style

Glomsda M., Prümer H. and Sieberg P. (2025). Estimation of Vehicle States Using a Cascaded Hybrid Estimation Method. In Proceedings of the 15th International Conference on Simulation and Modeling Methodologies, Technologies and Applications - Volume 1: SIMULTECH; ISBN 978-989-758-759-7, SciTePress, pages 126-134. DOI: 10.5220/0013646200003970

in Bibtex Style

@conference{simultech25,
author={Marvin Glomsda and Hendrik Prümer and Philipp Sieberg},
title={Estimation of Vehicle States Using a Cascaded Hybrid Estimation Method},
booktitle={Proceedings of the 15th International Conference on Simulation and Modeling Methodologies, Technologies and Applications - Volume 1: SIMULTECH},
year={2025},
pages={126-134},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0013646200003970},
isbn={978-989-758-759-7},
}

in EndNote Style

TY - CONF

JO - Proceedings of the 15th International Conference on Simulation and Modeling Methodologies, Technologies and Applications - Volume 1: SIMULTECH
TI - Estimation of Vehicle States Using a Cascaded Hybrid Estimation Method
SN - 978-989-758-759-7
AU - Glomsda M.
AU - Prümer H.
AU - Sieberg P.
PY - 2025
SP - 126
EP - 134
DO - 10.5220/0013646200003970
PB - SciTePress