نوع مقاله : مقاله پژوهشی
نویسندگان
1 دانشجو دکترا ، دانشگاه جامع امام حسین(ع)، تهران، ایران
2 دانشیار ، دانشگاه جامع امام حسین(ع)، تهران، ایران
3 استادیار ، دانشگاه جامع امام حسین(ع)، تهران، ایران
چکیده
کلیدواژهها
موضوعات
عنوان مقاله [English]
نویسندگان [English]
The classification modulation is operation between signal detection and demodulation in the physical layer and has been considered in the field of cognitive radio. so far no comprehensive research has been done in the DVB-S2X satellite signal modulation classification in the presence of GEO satellite channel effects. In this article, the conventional automatic modulation classification approach based on artifical features is combined with the evolutionary automatic modulation classification approach based on deep learning. Hiden and meaninful features of artificial features are automatically extracted by convolutional neural network. Also, the selection and optimization of artificial features is achieved by NSGAII
multi-objective genetic algorithm. The proposed algorithm performs automatic modulation classification of nine similar and nested modulations including high-order APSK modulations of the DVB-S2X signal. The proposed automatic modulation classification is evaluated with the presence of channel impeaments. The channel impeaments is effects of non-linear/non-ideal equipments including satellite transponder, ground station transmitter and the ground receiver terminal. performance of our manner is compared with other two methods. Using optimal artificial features despite the use of light convolutional neural network is strong performance in automatic modulation classification. In the signal to noise ratio above 5 dB, the overal accuracy is 100% in AWGN channel and 97% in GEO satellite channel. Also 100% classification accuracy was achieved for the value of signal to noise ratio which is lower than the value of signal to noise ratio of DVB-S2X standard.
کلیدواژهها [English]
)
