Spectrum Sensing Based on Cyclostationarity

Real time spectrum sensing with certain accuracy plays a key role in cognitive radio. Cyclostationary feature is used for spectrum sensing in this paper. Usually, cyclostationary feature detection requires high computation complexity, in this paper we analyze the performance of some frequencies and cycle frequencies for detection according to the licensed users' signal features, which reduce the complexity significantly. The best detection point is determined through simulation analysis on different detection points, and then we propose combination detection method using multiple detection points to obtain better performance. Results validate the effectiveness of the proposed method.