Research On Spectrum Sensing Technology In Cognitive Radios

Without generating interference to licenced transmissions, cognitive radio permits unlicenced services transmitting their signals in the licenced spectrum, which provides an alternative of effectively using the spectrum, and therefore, gradually becomes a hot research area. This paper focuses on exploring spectrum sensing, which is a critical technology of cognitive radio, and develops several novel sensing methods and strategies.Adopting cyclostationary detection as the basic technological route, this paper develops a detection method based on finding the optimum lag in cyclic autocorrelation, and extends this method to the detection of those primary user signals with multiple cyclic autocorrelation features. This method can achieve satisfying detection performance even under low SNR and interference conditions.Additionally, in order to effectively overcome the "shadowing effect", this paper proposes two strategies of cooperative detection. One is combing hard decision fusion rule and local cyclostationary detection results, while the other is using the developed soft data fusion rule synthesizing local test statistics.Moreover, in order to detect and recognize different secondary user signals, this paper adopts the specific cyclostationarity embedded OFDM signal, which was developed by previous people, as the secondary user signal, and deduces the corresponding theoretical cyclic autocorrelation equation of the special OFDM signal. From this equation, this paper applys the detection method of finding the optimum lag in cyclic autocorrelation into detecting the secondary user OFDM signals, and constructs a recognition system for it. As result, good detection and recognition performances are achieved in this paper.