Recursive State Estimation for Short-Term V2I Cluster-Delay Prediction
Scientific Journal of Technology · 21.10.2026 · 8 (10) · 92-99
DOI: 10.54691/8w78pb71 ↗Kurz gefasst
Fachartikel in der Zeitschrift Scientific Journal of Technology, online erschienen am 21.10.2026. Verfasst von Wanji Hou und 2 weiteren Autor:innen. Die Arbeit nennt 13 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY-NC 4.0).
Abstract (Originalsprache)
Short-horizon cluster-delay prediction must balance fluctuation suppression against responsiveness while accommodating newly observed clusters. We investigate a scalar local-level Kalman predictor that summarizes past observations recursively instead of requiring a fixed-length input window. Training-only trajectory validation selects its process-to-residual variance ratio; prediction precedes each observation update, and evaluation includes segment startup. Existing NYUSIM trajectories support direction holdouts, speed transfer, and a supplementary configuration-transfer evaluation with frozen parameters. At 25 m/s, the predictor reduces RMSE from 5.625 ns for persistence to 4.354 ns, closely matching a validated moving mean at 4.355 ns. The supplementary errors are 4.577 and 4.580 ns, respectively. Trend and innovation-adaptive extensions provide no consistent benefit. Startup errors and a segment-level counterexample delimit the results. These results characterize the accuracy and startup behavior of a compact recursive predictor within the evaluated stable segments.
Zitieren
APA
Hou, W., Lv, G., & Liu, S. (2026). Recursive State Estimation for Short-Term V2I Cluster-Delay Prediction. Scientific Journal of Technology, 8(10), 92-99. https://doi.org/10.54691/8w78pb71
Vancouver
Hou W, Lv G, Liu S. Recursive State Estimation for Short-Term V2I Cluster-Delay Prediction. Scientific Journal of Technology. 2026;8(10):92-99. doi:10.54691/8w78pb71
BibTeX
@article{Hou2026,
author = {Hou, Wanji and Lv, Gang and Liu, Shuang},
title = {Recursive State Estimation for Short-Term V2I Cluster-Delay Prediction},
journal = {Scientific Journal of Technology},
year = {2026},
volume = {8},
number = {10},
pages = {92--99},
doi = {10.54691/8w78pb71}
}
EUVIMED