Kurtosis based spectrum sensing for cognitive wireless cloud computing network

Spectrum sensing method for cognitive wireless cloud computing (CWC) network is very challenging since there are several different communication systems should be detected at very low SNR (as low as -22 dB). In this paper, we propose a kurtosis based spectrum sensing method which can be applied efficiently in such environment. The proposed method uses kurtosis estimation of received samples. Its value will be equal or close to 3 when only gaussian noise samples exist in the received signal. This kurtosis estimation's used to distinguish between the present or absent of primary signal by comparing with a predefined threshold. Simulation's done to evaluate its performance. Results show that the proposed method performs much better than energy detection especially at low SNR, even below -20 dB. It also gives benefit in much simple implementation for CWC network since it doesn't need knowledge of primary signal's parameters.