Fixed-Wing UAS Telemetry Benchmark (IDF_DS): 240 Flights for ML and Intelligent Trajectory Optimization

IDF_DS is an open-access corpus of fixed-wing UAS flight telemetry spanning 240 autonomous flights across two distinct circuits and two onboard architectures (SpeedyBee F405 running INAV, and Pixhawk 6X with a Jetson Orin NX companion running PX4). Missions were flown outdoors on repeatable, waypoint-guided trajectories and cover take-off, cruise, dynamic manoeuvres, and autonomous landing. Each log provides time-aligned, high-rate signals—including IMU kinematics, GNSS position/velocity, barometric altitude, airspeed (Pitot where available), flight-mode traces, actuator commands/states, and electrical metrics—enabling realistic modelling of flight dynamics, estimator behaviour, and sensor noise under varied conditions. To support diverse workflows, the dataset is distributed in two complementary layouts: Grouped Flights (one wide CSV per flight with all variables as columns) and Ungrouped flights (one folder per flight containing ~160 single-variable CSV files). Accompanying documentation specifies variable names, units, and logging rates, and flags known caveats (e.g., a small subset without calibrated Pitot that may require offset correction). IDF_DS is intended as a stable benchmark for comparative studies in trajectory tracking under GNSS degradation, wind-aware navigation and state estimation, anomaly detection, and energy-efficient planning, as well as for training and evaluating machine-learning pipelines for robust fixed-wing autonomy. The dataset is released under an open licence.