EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Robust Adaptive Extended Kalman Filtering for State of Charge Estimation Under Short-Term Current Measurement Loss

Zhi Ling, Junjiang Chen, Wenhao Guo, Suhong Wang, Chunxian Mo, Yongle Hu, Tuanfa Qin

Energies · 16.11.2026 · 19 (22) · 5536

DOI: 10.3390/en19225536 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Energies, online erschienen am 16.11.2026. Verfasst von Zhi Ling und 6 weiteren Autor:innen. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Guangxi Science and Technology Department.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Short-term current measurement loss disrupts charge integration and voltage prediction, challenging state of charge (SOC) estimation. We establish a strictly causal reconstruction and robust adaptive extended Kalman filter (RAEKF) procedure with voltage observations at every 1 Hz update. The physically consistent multirate interface preserves charge integration, polarisation history and endpoint current voltage response from 10 Hz reconstruction to 1 Hz filtering. SOC accuracy is assessed against simulated truth, while matched complete-current estimates provide a diagnostic of missing-current deviation and recovery. Paired comparisons, mechanism ablations and robustness tests characterise innovation protection and covariance adaptation, complemented by four newly generated synthetic holdout sessions. In these sessions, missing-interval true SOC mean absolute error (MAE) averages 0.141, 0.023 and 0.042 percentage points for persistence (Hold) with the extended Kalman filter (EKF), Hold–RAEKF and long short-term memory (LSTM) model C with RAEKF, respectively. The corresponding additional deviations are 0.131, 0.011 and 0.033 percentage points. The RAEKF suppresses larger average deviations in the 5 s and 20 s groups. LSTM-C has a lower additional SOC deviation than Hold in 5 of 48 RAEKF conditions; Hold is the stronger baseline overall. These findings connect multirate input physics, recursive error propagation and robust filter responses, informing reconstruction strategy selection and reliability assessment during short current measurement gaps.

Förderung

  • Guangxi Science and Technology Department

Zitieren

APA

Ling, Z., Chen, J., Guo, W., Wang, S., Mo, C., Hu, Y., & Qin, T. (2026). Robust Adaptive Extended Kalman Filtering for State of Charge Estimation Under Short-Term Current Measurement Loss. Energies, 19(22), 5536. https://doi.org/10.3390/en19225536

Vancouver

Ling Z, Chen J, Guo W, Wang S, Mo C, Hu Y, et al. Robust Adaptive Extended Kalman Filtering for State of Charge Estimation Under Short-Term Current Measurement Loss. Energies. 2026;19(22):5536. doi:10.3390/en19225536

BibTeX
@article{Ling2026,
  author = {Ling, Zhi and Chen, Junjiang and Guo, Wenhao and Wang, Suhong and Mo, Chunxian and Hu, Yongle and Qin, Tuanfa},
  title = {Robust Adaptive Extended Kalman Filtering for State of Charge Estimation Under Short-Term Current Measurement Loss},
  journal = {Energies},
  year = {2026},
  volume = {19},
  number = {22},
  pages = {5536},
  doi = {10.3390/en19225536}
}
Kostenlos starten →