Research On Cooperation Of Better Cognitive Users Spectrum Sensing Technology

The spectrum sensing is the core technology of the cognitive radio(CR).Cognitive radio communication system must have spectrum sensing capability, andthat CR users need accurately sense whether there are idle bands so as to use thesebands for communicating. At the same time, CR users must momentarily monitorwhether authorized users need access to the band, this makes the CR user to promptlywithdraw for avoiding causing harmful interference to authorized users. Therefore,the spectrum sensing need to fulfill two functions: at first, sensing spectrum hole andmaking full use of all spectrum opportunities; the second, detecting the emergence ofauthorized users to avoid causing interference to authorized users. Accurate sensingidle spectrum and timely detection the emergence of authorized users, these need tohave sufficient time to complete; in addition, in order to effectively transfer data, thesystem need guarantee for spectrum detection time as short as possible, so as toimprove the utilization of the spectrum resources better. Apparently the above twofunctions are mutually contradictory. So, the core issue of spectrum sensingtechnology is how to achieve a reasonable compromise between detectioneffectiveness and detection accuracy.This thesis firstly studies the theoretical knowledge of the cognitive radio as wellas the classic spectrum sensing algorithms, then focuses on cooperative spectrumsensing algorithm, and points out that the cooperative detection among users canconquer lots of the limitations of single-user detection so as to improve the systemdetection performance. Cooperation spectrum sensing firstly reports the local decisionmessage of each user to the fusion center, then the fusion center makes the finaldecision according to the most fitting fusion rule and feed back the decision results toCR users. Fusion rule is mainly divided into OR rule and AND rule. The traditionalcooperative spectrum sensing refers to all the cognitive users are involved incooperation. By analyzing the performance of traditional cooperative spectrumsensing under low Signal-to-Noise, this thesis points out the limitations that all thecognitive users are involved in cooperation, then presents with higher signal-to-noiseratio cognitive users to participate in cooperation. Then how to choose these users,namely adopting what kind of user selection algorithm becomes the key of achieving this idea. Further，this thesis proposes the better user selection algorithm anddetermines the optimal threshold by the simulation. From the simulation result, thedetection performance of this method compared with traditional cooperative detectionalgorithm has apparent improvement. Therefore, compared to all users involved in thetraditional cooperation, the selected cooperation among cognitive users can betterimprove the sensing capability of the system.