An open benchmark dataset for machine learning and intelligent trajectory optimization in fixed-wing unmanned aerial...

Area of Science:

• Robotics and Control Systems
• Aerospace Engineering
• Data Science

Background:

• Intelligent unmanned aerial systems (UAS) require robust datasets for research and development.
• Existing datasets may lack the comprehensive, high-resolution telemetry needed for advanced autonomous applications.

Purpose of the Study:

• To present a novel, open-access telemetry dataset for intelligent fixed-wing UAS.
• To support research in autonomous flight, navigation, and system performance analysis.

Main Methods:

• Collected 240 autonomous flight logs from two distinct onboard architectures (SpeedyBee F405/INAV and Pixhawk 6X/Jetson Orin NX/PX4).
• Recorded synchronized multi-sensor telemetry (IMU, GNSS, barometric altitude, actuators, power) at high temporal resolution.
• Included diverse flight phases: take-off, cruise, dynamic maneuvers, and autonomous landing.

Main Results:

• The dataset provides detailed, annotated data for 240 autonomous missions.
• Enables realistic modeling of flight dynamics, estimator behavior, and sensor noise.
• Facilitates benchmarking for trajectory tracking, anomaly detection, and energy-optimized mission planning.

Conclusions:

• The dataset is a valuable resource for advancing research in intelligent fixed-wing UAS.
• Supports development and testing in challenging operational environments, including contested and safety-critical scenarios.
• Promotes innovation in autonomous navigation and control strategies for unmanned aerial systems.