In this paper, new algorithms for spectrum sensing in cognitive radio based on higher order cumulants and kurtosis are proposed. The cumulants represent statistical signal processing based on pattern recognition for signals of different structure, and has low implementation complexity. Kurtosis statistics are a well-known technique for testing the Gaussianity feature of the signals. Under the assumption that a detected signal can be modelled according to an autoregressive model, noise variance is estimated from that noisy signal. The simulation results show that spectrum sensing algorithms based on the estimated normalised values of joint higher order cumulants (of fourth and sixth orders) and kurtosis are reliable for a wide range of signal-to-noise ratio environments. In order to improve performances of the spectrum sensing, the combination of these statistics tests into unique one statistic test is proposed. Simulation results have verified improvement of the performances.

No abstract

No abstract

No abstract

No abstract