EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Learning Calibrated Uncertainty Margins for an Explainable 4D-Radar UAV Warning Policy

Le Huang, Jiawen Dai, Jiaojiao Liu, Jie Cao, Yongjun Cai, Jie Bai, Bangji Zhang, Zheshuo Zhang

Information · 05.11.2026 · 17 (11) · 1105

DOI: 10.3390/info17111105 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Information, online erschienen am 05.11.2026. Verfasst von Le Huang und 7 weiteren Autor:innen. Die Arbeit nennt 37 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: National Key Research and Development Program of China, Key Research and Development Program of Jiangxi Province, 2025 Hangzhou Innovation and Entrepreneurship Program for High-Level Overseas Returnees.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Reliable collision warning from 4D millimeter-wave radar is hindered by the evaluation of deterministic distance–speed boundaries on uncertain track states. Directly learning the alarm decision may improve flexibility, but obscures safety-critical logic. To reconcile robustness with interpretability, this paper develops a hybrid uncertainty-aware warning framework in which a lightweight multilayer perceptron (MLP) estimates deviations from nominal radar tracks, whereas the final binary decision remains governed by explicit geometric and kinematic conditions. Validation-residual calibration transforms the instantaneous estimates into empirical upper margins, and a projection-bounded leakage recursion preserves recent uncertainty during brief observation gaps. The margins act only on the time-based entry branch; the hard-distance override, motion-direction gate, confirmation logic, and release rule remain deterministic. The framework is evaluated using 36 fixed-radar, single-reflector physical recordings and 216 trial-isolated controlled-degradation realizations derived from them. The calibrated margins attain 95.24% range-error coverage and 94.04% speed-error coverage relative to the nominal tracks. Among 198 realizations with a reference entry, robust-margin methods reduce late or missed warnings from 15 to 2. At this protected operating point, the calibrated MLP decreases alarm disagreement from 4.94% for a fixed margin to 3.90%, compared with 4.16% for a calibrated linear model. The results identify a condition-dependent margin-allocation benefit while retaining an auditable warning policy for the tested recordings.

Förderung

  • National Key Research and Development Program of China
  • Key Research and Development Program of Jiangxi Province
  • 2025 Hangzhou Innovation and Entrepreneurship Program for High-Level Overseas Returnees

Zitieren

APA

Huang, L., Dai, J., Liu, J., Cao, J., Cai, Y., Bai, J., Zhang, B., & Zhang, Z. (2026). Learning Calibrated Uncertainty Margins for an Explainable 4D-Radar UAV Warning Policy. Information, 17(11), 1105. https://doi.org/10.3390/info17111105

Vancouver

Huang L, Dai J, Liu J, Cao J, Cai Y, Bai J, et al. Learning Calibrated Uncertainty Margins for an Explainable 4D-Radar UAV Warning Policy. Information. 2026;17(11):1105. doi:10.3390/info17111105

BibTeX
@article{Huang2026,
  author = {Huang, Le and Dai, Jiawen and Liu, Jiaojiao and Cao, Jie and Cai, Yongjun and Bai, Jie and Zhang, Bangji and Zhang, Zheshuo},
  title = {Learning Calibrated Uncertainty Margins for an Explainable 4D-Radar UAV Warning Policy},
  journal = {Information},
  year = {2026},
  volume = {17},
  number = {11},
  pages = {1105},
  doi = {10.3390/info17111105}
}
Kostenlos starten →