HisarMod: a new challenging modulated signals dataset

In order to increase the diversity in signal datasets, we create a new dataset called HisarMod, which includes 26 classes and 5 different modulation families passing through 5 different wireless communication channel. During the generation of the dataset, MATLAB 2017a is employed for creating random bit sequences, symbols, and wireless fading channels.

The dataset includes 26 modulation types from 5 different modulation families which are analog, FSK, PAM, PSK, and QAM. In the dataset, there are 1500 signals, which have the length of 1024 I/Q samples, for each modulation type. To make HisarMod similar to RadioML2016.10a for a fair comparison, there are 20 different SNR levels in between -20dB and 18dB. As a result, the dataset covers totally 780000 signals. When generating signals, oversampling rate is chosen as 2 and raised cosine pulse shaping filter is employed with roll-off factor of 0.35.

Mots clés

• Modulation Classification
• Deep Learning
• cognitive radio