The Simulation And Analysis Of Spectrum Sensing Based On Higher Order Cumulants

Cognitive radio has emerged as one promising technology for solving current spectrum resource scarcity. A cognitive user is able to sense the outside world and to learn the information of surroundings. It can make the communication more reliable while improving the utilization of radio spectrum by changing real time operation parameters to accommodate the internal to the outside world. Spectrum sensing is a primary task in cognitive radio. After analyzing some spectrum sensing methods, this dissertation proposes a new spectrum sensing method based on higher order cumulants. The spectrum sensing methods, such as energy detection, can not get further information of radio spectrum. So this dissertation verifies spectrum sensing method based on higher order cumulants by simulating and focuses on modulation classification, which make cognitive radio more intelligent.This dissertation firstly gives a basic introduction of cognitive radio, including its definition, main functions, important technologies and so on. With regard to sensing spectrum techniques, this dissertation systematically investigates some common spectrum sensing techniques: matched filtering, energy detection, cyclostationary feature detection and other present spectrum sensing methods.In addition, this dissertation analyses the spectrum sensing method based on higher order cumulants. According to the analysis of higher order cumulants theory, it is better to adopt bispectrum detection method considering performance and calculation. With this method, this dissertation detects VSB and DS-SS/BPSK signals under different signal to noise ratios through simulating. The result shows that bispectrum detection method can detect weak signals even at a low signal to noise ratio.Besides the usage of radio spectrum, cognitive users should also know more information such as modulation schemes of signals. At present, the calculation of modulation classification based on higher order cumulants is usually large. It is difficult for engineering reality. This dissertation analyzes and combines the peak energy of bispectrum diagonal slice and spectrum analysises to classify the modulation schemes of signals. Through simulating, the result proves that the method is efficient.Finally, the dissertation researches hardware implementation of cognitive radio including the spectrum sensing hardware structure in some drafts, then proposes an improved sensing structure. This structure decreases sensing range and improves the performance of system with some prior knowledge.