More Datasets

AERPAW RF Sensor Measurements with UAV July 2024

2025
2024, 3.3 GHz, Drone, Localization, Rural
GE2: Channel Sounder, Keysight RF Sensor, KRSE3: TDOA-Based Source Localization

Lead Experimenter

Cole Dickerson, North Carolina State University

Link to Dataset

The dataset contains TDOA localization estimates, ground-truth UAV positions, and binary LOS/NLOS indicators for each tower involved in the TDOA estimation. The dataset is available via [Dryad].

Equipment and Software Used

Keysight N6841A RF sensors on AERPAW towers LW2–5, Keysight N6854A Geolocation Software, GNU Radio, and an AERPAW UAV equipped with an SDR portable node.

Description

The experimental efforts for which this data was collected aimed to evaluate the performance of Time Difference of Arrival (TDOA)-based 3D localization of UAVs in real-world environments. A UAV equipped with a software-defined radio transmitted RF signals at a central frequency of 3.32 GHz, using varying bandwidths (1.25 MHz, 2.5 MHz, and 5 MHz) and altitudes (40 m, 70 m, and 100 m) over controlled flight trajectories. Four Keysight N6841A RF sensors were deployed on fixed towers to capture RF signals and process them using TDOA algorithms for position estimation. The experiments were conducted under mixed line-of-sight (LOS) and non-line-of-sight (NLOS) conditions with the additional use of ray-tracing simulations to analyze LOS/NLOS effects. These efforts sought to understand how altitude, bandwidth, and environmental factors impact UAV localization accuracy.

Publications

  1. C. Dickerson, S. Masrur, J. Dickerson, O. Ozdemir, and I. Guvenc, "Impact of Altitude, Bandwidth, and NLOS Bias on TDOA-Based 3D UAV Localization: Experimental Results and CRLB Analysis," Proc. IEEE ICC Workshops, Wkshp. on Positioning and Sensing over Wireless Netw., Montreal, Canada, June 2025, to appear.

Representative Results

The figures collectively illustrate the UAV’s 3D ground truth trajectory, the comparison between estimated and actual positions in 2D, and spatial error patterns. The first figure maps ground truth positions with corresponding error levels and tower locations. The second overlays error magnitudes on the 3D trajectory using color coding (green for errors <100 m, red for ≥100 m). The third compares estimated and ground truth coordinates in 2D, including tower locations. The final figure presents the UAV’s 3D flight path with annotated start and end points.