Um método para discriminação entre PSK e FSK utilizando estatísticas de ordem superior

This paper presentes a new method to discriminate between phase and frequency digital modulations. The algorithm is able to identify the class of an M-PSK or M-FSK modulated signal, for any modulation order M. The proposed method is based on machine learning, using higher-order statistics of the instantaneous frequency of the baseband received signal as features, presenting a good tradeoff between complexity, a priori knowledge and performance, compared to the existing methods