Energy detection method enhanced by autocorrelation

Abstract

a hybrid algorithm for primary user detection in spectrum sensing is proposed in this article, and its performance for different types of primary signals is evaluated and compared. The hybrid algorithm is a combination of the energy detection and the cyclostationarity-based methods. In the majority of related works, the input signal for the energy detection is modeled with a lognormal distribution without considering the cyclostationarity of signal. However, primary signals may have other behaviors with different cyclostationarity levels. Those behaviors depend on the modulation scheme used and the traffic statistics of the primary user. Therefore, the hybrid method is evaluated considering several input signal models, which provide different levels of cyclostationarity. Additionally, a measured signal is used for evaluating the energy detection and the proposed method. The performance of the proposed method shows a detection gain in relation to the energy detection. The proposed method takes advantage of the higher accuracy of cyclostationarity method and the simplicity of energy detection. The obtained results are important in the investigation of a more generic and more accurate method for detecting primary users in systems such as wireless sensor networks.