RadioML2016.10A\/10B\/10C

RadioML 2016.10A, released by DeepSig as part of the RadioML project, is an open-source synthetic radio signal dataset introduced at the 6th GNU Radio Conference in 2016. It is an improved version of the earlier 2016.04C release, offering a cleaner and more standardized format.The dataset encompasses 11 modulation types (e.g., 8PSK, WBFM, AM-DSB) across SNRs ranging from -20dB to 18dB (2dB steps), with 1,000 samples per modulation-SNR pair (resulting in a total of 220,000 signals). The dataset emulates real-world impairments, including multipath fading, local oscillator offsets, and frequency-selective fading. This configuration provides a rigorous benchmark for robustness testing under practical wireless channel distortions.