An Improved Algorithm Based on First-order Cyclostationarity for Recognition of M-FSK Signals

Abstract

This paper proposes an improved algorithm based on Piecewise Auto Correlation Accumulation (PACA) and first-order cyclostationarity for recognition of M-ary frequency shift keying (M-FSK, M = 2/4/8) signals in communication systems. Theoretical analysis indicates that the number of first-order Cycle Frequencies (CFs) is equal to the modulation order of M-FSK signals. Therefore, by estimating the first-order cyclic moment of the autocorrelation of M-FSK signals and detecting the number of CFs, the modulation order of M-FSK signals can be identified. Based on its multicarrier characteristics, the proposed algorithm does not require a priori information and is not significantly influenced by factors such as timing delay and frequency offset. Further, PACA suppresses noise, and thereby helps to improve the antinoise performance of the algorithm’s modulation recognition. The results of simulations conducted confirm the efficacy of the proposed algorithm.