Article: Time series data-driven UAV sensor attack detection: an adaptive graph-time-frequency hybrid approach Journal: International Journal of Automation and Control (IJAAC) 2026 Vol.20 No.7 pp.1 - 25 Abstract: To address the limitations of existing methods in dynamic spatial learning, joint time-frequency analysis, and drift attack detection for UAV sensor security, this paper proposes the graph time-frequency mixed anomaly detection (GTF-MAD) model. Its core innovations include an adaptive clustering mask graph attention network (ACM-GAT) for dynamic sensor correlation learning, a time-frequency dual-stream cross-processing path (TFDSCPP) for deep multi-domain feature fusion, and the TSM-EWMA trend detection method for rapid identification of drift attacks. Experiments on a quadrotor platform confirm GTFMAD's superior performance, achieving a peak F1-score of 99.71% for bias attacks and reducing drift attack detection latency by 67.2% (from 119 to 39 steps) compared to traditional methods. The model offers a reliable and high-precision solution for real-time UAV sensor security. Inderscience Publishers - linking academia, business and industry through research

Title: Time series data-driven UAV sensor attack detection: an adaptive graph-time-frequency hybrid approach

Authors: Junfeng Chen; Yuhang Zhou; Xingsi Xue

Addresses: College of Artificial Intelligence and Automation, Hohai University, Changzhou, 213200, China ' College of Artificial Intelligence and Automation, Hohai University, Changzhou, 213200, China ' Fujian Provincial Key Laboratory of Big Data Mining and Applications, Fujian University of Technology, Fuzhou, Fujian, 350118, China

Abstract: To address the limitations of existing methods in dynamic spatial learning, joint time-frequency analysis, and drift attack detection for UAV sensor security, this paper proposes the graph time-frequency mixed anomaly detection (GTF-MAD) model. Its core innovations include an adaptive clustering mask graph attention network (ACM-GAT) for dynamic sensor correlation learning, a time-frequency dual-stream cross-processing path (TFDSCPP) for deep multi-domain feature fusion, and the TSM-EWMA trend detection method for rapid identification of drift attacks. Experiments on a quadrotor platform confirm GTFMAD's superior performance, achieving a peak F1-score of 99.71% for bias attacks and reducing drift attack detection latency by 67.2% (from 119 to 39 steps) compared to traditional methods. The model offers a reliable and high-precision solution for real-time UAV sensor security.

Keywords: UAV; unmanned aerial vehicle; sensor attack; anomaly detection; GNN; graph neural network; time-frequency analysis; trend detection.

DOI: 10.1504/IJAAC.2026.153751

International Journal of Automation and Control, 2026 Vol.20 No.7, pp.1 - 25

Received: 22 Sep 2025
Accepted: 19 Nov 2025
Published online: 22 May 2026 *

Title: Time series data-driven UAV sensor attack detection: an adaptive graph-time-frequency hybrid approach

Keep up-to-date