An Improved Algorithm for Blind Modulation Recognition of Signals in Cognitive Radio ⋆

An improved algorithm for blind modulation recognition of signals in cognitive radio (CR) is proposed in this paper, which uses the number of sinusoid component as a feature to classify FSK, BPSK, MSK and QPSK modulation schemes. Firstly, the received signal is preprocessed by short time filtering to improve the processing signal-to-noise ratio (SNR). Secondly, nonlinear operations are applied to signals, and then signals of different modulation schemes can manifest themselves in the number of sinusoid component, which is equivalent to a line in the spectrum. Simulation results show that the average success identifying rate of the proposed algorithm is beyond 90% when the SNR is no less than -1dB.