Distributed Spectrum Sensing In Cognitive Radios

In recent years, with the increasing requirement of communication efficiency and quality and the lack of spectrum resource, the traditional method of spectrum detection strategies and static spectrum allocation which has caused low spectrum utilization rate can’t satisfy the demand for high quality of communication network services. As a promising method to alleviate the scarcity of spectrum resources, cognitive radio has become a hot research field of wireless networks developed rapidly. Spectrum sensing technology is one of the main research directions and a key component of cognitive radio, and can solve the problem of low spectrum utilization rate. Thus, it is necessary to put forward a novel and efficient spectrum detection method.In this paper, firstly, the concept of cognitive radio is proposed, and the research background is also introduced, as well as the current research status in china and abroad. The research direction is also been presented.Secondly, various spectrum detection technologies are introduced. It could be classified into two parts:single spectrum sensing and cooperative spectrum perception. The first part introduces single spectrum sensing methods, such as energy detection, matched filtering, and cyclostationary feature detection, and advantages and disadvantages of these methods are also compared. The second part introduces cooperative spectrum perception technology, and the two common methods-centralized cooperation detection and distributed cooperation detection are presented in detail. Then, the reliability of cooperation detection is analyzed.Thirdly, spectrum sensing based on nonnegative matrix factorization is proposed. As the basic method of single spectrum sensing, nonnegative matrix factorization and the application in spectrum sensing is introduced detailed. Through simulation, nonnegative matrix factorization detection is contrasted with two traditional algorithms, i.e., energy detection and cyclostationary feature detection.Finally, after analyzing distributed cooperative spectrum sensing theory, nonnegative matrix factorization detection which has the function of parallel detection is applied to distributed cooperative spectrum sensing, as the basic method of local sensing detection. Through the establishment simulation system, nonnegative matrix factorization cooperative sensing is checked, and the advantage of it is contrasted with energy detection cooperative sensing.