Joint classification and parameter estimation of M-FSK signals for cognitive radio

Spectrum sensing and awareness constitute key functionalities of a cognitive radio (CR), and encompass signal detection, classification, and blind parameter estimation. This paper proposes a novel algorithm for the tone frequency spacing estimation and joint classification of M-ary frequency shift keying (M-FSK) signals in a fading environment. The proposed algorithm relies on the number and position of the first-order cycle frequencies (CFs), and requires neither recovery of the carrier and symbol timing, nor the estimation of channel parameters or signal and noise powers. Receive spatial diversity is exploited to enhance the classification and estimation performance of the algorithm. The simulation results for the algorithm performance confirm its effectiveness.