The Joint Spectrum Sensing Selective Strategy In Cognitive Radio Networks

Cognitive radio, whose main propose is to realize dynamic spectrum access based on spectrum sensing, is the key technology to resolve deteriorated poor spectrum utilization problem. In order to improve the spectrum utilization quality, the dynamic access uses unfixed spectrum allocation mode. The spectrum resource dynamic accessed by cognitive users are authorized to primary users, thus spectrum sensing requires strict quality which the cognitive users must obtain accurate channel information to lead the influence to primary users within tolerable range.Within those spectrum sensing models, the model based on energy detection with simple and practical quality has become a research hotspot. Under the complex wireless communication environment, compared with single node detection mode, joint spectrum sensing mode is widely researched which can improve the detection accuracy. In this dissertation, the selection strategy of detection nodes under the complex wireless communication environment is discussed with information fusion mechanism. On the other hand, the research status of existing joint spectrum sensing modes is analyzed.Joint spectrum sensing mode distinguishes the quality of nodes'received signal, including soft selection mechanism and hard selection one, in which the former uses weighted mode to distinguish signal quality; and the latter by utilizing good quality nodes for information fusion. This dissertation concerns hard selection mechanism according to channel station information based on information fusion of key nodes. Therefore, the main work of node selection strategy is to acquire selection reference. Within the current design models, signal to noise ratio (SNR) is considered as reliable selection reference of key nodes. Based on this propose, the strategy of choosing high SNR nodes as key nodes has been researched by many models, but those models face calculation problem, when the detection nodes work in low SNR environment. To solve this problem, this work introduces a high performance joint spectrum sensing model based on key nodes instead of high SNR. The selection strategy of this mode is based on"fact"reference, which can avoid SNR calculation problem. Additionally, the time window mechanism (TWM) is designed to select key nodes to generate"fact"reference, and corresponding information fusion is rethought too. According to different information forms, two selection models which are hard fact selection model and soft fact selection model respectively are utilized. Finally, the simulation under AWGN channel condition is studied, which compared with the model combining all nodes with indistinct selection.The simulation result shows that new models can achieve better precision quality without using SNR reference. Otherwise, the simulation demonstrates that dealing with more information, soft fact selection model has better quality than hard fact selection model.