Multi-task Learning Dataset for Automatic Modulation Classification and DOA Estimation

A synthetic signal dataset of 12 differentmodulations (including PSK, QPSK, 8PSK, QFSK, 8FSK,16APSK, 16QAM, 64QAM, 4PAM, LFM, DSB-SC, and SSBSC) with different DOAs (discrete angles ranging from -60°to 60°with the step size of 1°) is generated using MATLAB2021a. Regarding the signal model configuration for the datageneration, we specify a uniform linear antenna array of M = 5 elements to acquire incoming signals having N = 1024envelope complex samples, thus conducting an I/Q data arrayof size 1024 × 2 × 5. To model real-world phenomena inwireless communication, various impairments are considered:the Gaussian noise in the range [-10, 20] dB (step size of 1dB), the multi-path propagation with a random number of NLOSsignals in the range of [1, 10], the propagation attenuation αpand delay τp randomly distributed in the range [-50, -1] dBand [1, 3000] ns, respectively. The dataset contains 450120signals covering 121 DOA classes, 12 modulation classes at 31SNR levels, where each signal has 1024 samples to form intoan input data array with the size of 1024x2x5.