LoRadar: A Raw IQ Dataset of Satellite-Ground LoRa Communication Signals

LoRadar is a large-scale satellite-to-ground LoRa signal dataset collected using a HackRF SDR platform and a ground-based terminal. It contains 181 raw IQ signal bursts, each representing an uplink LoRa packet transmitted from an IoT terminal to a low-Earth-orbit (LEO) satellite. The total dataset size is approximately 127 GB.

All signals are stored in `.bin` files using complex64 format (interleaved float32 I/Q samples). The dataset spans the 399–403 MHz frequency range with a uniform sampling rate of 4 MHz. Each packet corresponds to a different configuration of LoRa physical-layer parameters, including center frequency, bandwidth (typically 125 kHz or 250 kHz), and spreading factor (ranging from SF7 to SF12). Signals were collected during real satellite overpasses under varying SNR and Doppler conditions