UAV-SEAD: State Estimation Anomaly Data for PX4 Flight Controllers

UAV-SEAD provides real-world multivariate time-series telemetry for state estimation anomaly detection in Unmanned Aerial Vehicles (UAVs), released by Aykut Kabaoglu and Sanem Sariel in 2026. The data supports research into Fault Detection and Identification (FDI) specifically for PX4-based flight systems.

Use Cases

• Training anomaly detection models on multivariate PX4 telemetry logs
• Developing Fault Detection and Identification (FDI) algorithms for state estimation
• Benchmarking flight controller stability during anomalous state transitions

Strengths

• Contains real-world flight telemetry rather than purely synthetic data
• Specifically targets state estimation anomalies in PX4-based systems
• Released under a CC BY 4.0 license for open research

Limitations

• Specific column headers and feature counts are not explicitly listed in the provided metadata
• Environmental flight conditions and specific sensor noise levels are not specified

Provenance

Source

UAV-SEAD: State Estimation Anomaly Dataset for UAVs (2026)

Collection Method

Real-world flight sensor logging

Freshness

Last updated February 2026

Geography

United States