The widespread use of drones across industries has raised concerns about security threats. To provide a publicly available data set for developing robust drone detection and classification systems, we present a multi-sensor dataset composed of time-synchronized signals from Frequency-Modulated Continuous Wave (FMCW) radar, Continuous Wave (CW) radar, and Radio Frequency (RF) receiver. The dataset includes raw and processed signals collected from four different commercial drones and one non-drone target, and signals for different sensors are collected simultaneously. Measurements were taken across distances ranging from 2 to 30 meters in 2-meter intervals, with repeated trials under controlled conditions to ensure consistency. Unlike existing datasets that rely on a single sensing modality, our dataset enables direct comparison and fusion of multi-sensor signals. It contains both raw signals and processed representations-range-Doppler maps, Doppler spectrum, and power spectral densities-facilitating signal-level and image-based analysis. This time-synchronized dataset offers a reliable resource for developing and evaluating multimodal sensor-fusion strategies, distance-aware classification models, and advanced Artificial Intelligence-based detection algorithms, including supervised and self-supervised learning.
Han et al. (Tue,) studied this question.