Distributed Cooperative Spectrum Sensing in Cognitive Radio Networks with Adaptive Detection Threshold

This paper proposes a cooperative sensing algorithm based on distributedfusion strategy and maintained probability of false alarm for cognitive radio. It furtherintroduces a reporting strategy that discusses how cooperative sensing in distributedmanner can select among possible candidates in order to reduce bandwidthrequirement. We adopted a dynamic distributed architecture for cooperative sensingbased on the link quality and found condition on the channel quality for cooperation tobe beneficial. Using probability of detection, and BER metrics we evaluated theperformance improvement of distributed cooperation over direct cooperation and noncooperativesensing. We used analytical formulation with possible candidate selectioncriteria to investigate and maximize the cooperation gain. By employing suchdistribution and selection technique, the reporting error due to the fading channel isreduced. Results show that the method effectively improve performance of sensing, itincrease the probability of detection up to 0.9 at <0.1 probability of false alarm.Sensitivity requirement is reduced with network scale and the number of nodesparticipate in decision fusion is reduced about 42% at probability of false alarm 0.1.ROC curve has obvious improvement compared with existing methods