AeroTSBoost: Temporal-Statistical Boosting for Real-World UAV Telemetry Anomaly Mining
Abstract
Mining anomalies from unmanned aerial vehicle (UAV) state-estimation logs is challenging because failures are sparse, temporally structured, and distributed across heterogeneous PX4 telemetry streams with variable sensor availability and missing values. We present AeroTSBoost, a temporal-statistical boosting framework for real-world UAV telemetry anomaly mining. AeroTSBoost aligns multivariate flight logs, converts each window into deterministic descriptors that capture distributional shifts, quantile structure, endpoint drift, local dynamics, and lag correlation, and trains a class-balanced LightGBM detector. On UAV-SEAD, AeroTSBoost achieves the strongest AUPRC among evaluated classical, supervised tabular, neural reconstruction, recurrent, Granger-causality-based, and frequency-domain baselines. Across five seeds, it reaches AUPRC and threshold-swept event F1, improving AUPRC by 5.79 absolute points over the strongest non-AeroTSBoost baseline. Under purged chronological and leave-log-out protocols, it remains the best AUPRC method, reaching and , respectively. On related ALFA fixed-wing UAV fault logs, AeroTSBoost reaches leave-sequence-out AUPRC, ahead of RandomForest () and moments-only (). These results show that deterministic temporal-statistical representations remain highly competitive for sparse anomaly mining in operational cyber-physical telemetry.
Cite
@article{arxiv.2605.25639,
title = {AeroTSBoost: Temporal-Statistical Boosting for Real-World UAV Telemetry Anomaly Mining},
author = {Junhao Wei and Haochen Li and Yanxiao Li and Yifu Zhao and Dexing Yao and Baili Lu and Xudong Ye and Sio-Kei Im and Yapeng Wang and Xu Yang},
journal= {arXiv preprint arXiv:2605.25639},
year = {2026}
}