Cognitive Radio Frequency Domain Sparse Signal Detection And Statistic

As an intelligent spectrum sharing technology, cognitive radio can detect the electromagnetic wireless environment and adaptively change its system operating parameters to effectively use the free spectrum relying on particular spectrum sensing technologies. However, for some broadband or high-frequency signals,typical spectrum signal detection methods are restricted by the Nyquist sampling theorem, also caused much waste of sampling resources. The introduction of compressed sensing thinking exactly targeted to solve this problem. When a signal is detected as a sparse signal, it collected a small number of values which is projected from high-dimensional space to a lower dimensional space, to achieve recovery of the signal by certain reconstruction algorithms.This paper firstly introduces some typical spectrum sensing techniques such as energy detection, matched filtering detection, cyclo-stationary feature detection, as well as their respective advantages and disadvantages for different occasions. Then focuses on the basic framework of compressed sensing and signal reconstruction algorithm, selecting three of the most common greedy iterative algorithm(matching pursuit, orthogonal matching pursuit and Segmented Matching Pursuit algorithm) for simulation analysis. However, it doesn't need to completely and accurately reconstruct the original signal in such signal detection problem, merely determine the spectral position of the primary user or interference signal in frequency domain through certain partial reconstruction algorithm.Therefore, the core of this paper is using the idea of compressed sensing to partially reconstruct the detected spectrum, as well as the establishment of a cognitive model based on compressed sensing signal detection theory. Then researches the partially reconstruction detection algorithm upon the three basic completely reconstruction algorithm mentioned above, which enables the cognitive users to signal sampling using far less than the Nyquist frequency requirements while ensuring the success rate of the detection and immensely saving the sampling resources. In particular, the simulation experiment proves the StOMP detection algorithm can guarantee a higher range of threshold without interfering of primary users communication, which also ensures the success rate of signal detection. Finally, to the question of existing spectrum analysis haven't sufficiently considered the issue of the spectrum resources reliability, we propose a measurement approach of the spectrum resources reliability based on channel availability and stability as a supplement to theexisting spectrum analysis content. It improves the analytical methods, making the results of spectral analysis more comprehensive and more convincing.