Research On Wireless Spectrum Environment Cognition Technologies

With the innovation of communication technology and unceasing increase in socialdemands, wireless communication is playing a significant role. Meanwhile, however, itis being confronted with many issues and contradictories, such as the contradictorybetween fixed static allocation methods in traditional wireless spectrum and currentdemands for flexible wireless spectrum, the contradictory between networkheterogeneity due to various communication standards and current demands for fusionof access and services, and the contradictory between the scarcity of wireless frequencyresources and the low efficiency in wireless spectrum usage.In order to tackle these issues and contradictories, Cognitive Radio has beenproposed and studied. It is a new kind of smart wireless communication technologieswhich can boost significantly the usage efficiency of wireless spectrum by the way thattakes advantages of detecting, analyzing, learning, reasoning and planning the wirelessspectrum environment without disturbing primary users. On the basis of relevant resultsof wireless spectrum environment, it is able to share wireless spectrum resources byadjusting its own transmission parameters in time, space, frequency, modulationmethods and so forth. Current wireless spectrum environment is a complex environmentweaved by multiple standards, various modes and heterogeneous networks.Consequently, several crucial technologies of wireless networks such as environmentcognition, independent decision-making, and network reconstruction have to becombated to transfer theories of Cognitive Radio to practical usage, in which Cognitivetechnology on wireless spectrum environment is a premise and crucial point forachieving Cognitive Radio. One task of cognition on wireless spectrum environment isthe cognition of distinguishing accurately and quickly the busy and idle status ofspectrum environment. Another task of radio spectrum environmental sensing is toprovide sufficient information support for secondary users to achieve subsequenttransmission.Having deeply studied these two tasks, the main innovation of our paper is focusedon the following four aspects:First, targeting at the contradictory between fusion performance and throughput in the research of credible fusion of cognitive radio networks, this paper proposes andexplores a reliable cooperative spectrum sensing algorithm based on double thresholdand D-S evidence theory from authors’ spectrum sensing research. In this algorithm,double threshold energy detection is used by local secondary users to enhance thereliability of detection. The results of double threshold energy detection are classifiedinto two categories: in one category the results are decided locally and then decidedresults are sent to fusion center in the form of0and1; in the other category in whichvalues between the two thresholds cannot be decided directly, the trust function is sentto fusion center instead. In fusion center, the second category of soft information data isprocessed according to D-S evidence theory first and results are combined with the firstcategory results to make the final decision. As for fusion center, this algorithm isequivalent to select nodes first, and thus reduce the burden on network transmission andthe burden on fusion center for computation.Secondly, targeting at the issue of false alarm probability in proposed DIDSalgorithm DADS algorithm, this paper focus on the effect of weighting coefficient thetotal deciding performance in fusion center, and propose an improved algorithm thatintroduces weighting coefficient into sensing results of users, which improves the totalperformance and reliability of the system. Simulation results indicate that the detectionprobability can maximally improve30%in our algorithm compared to other single nodeD-S evidence theory and traditional cooperative cognition decisions based on “or” or“and” decision, while the performance of the receiver operating performance curve(ROC) improve3%compared to single D-S evidence theory which is regarded asrelative good algorithm.Thirdly, targeting at the difficulty of separating mixed signals of wideband system, analgorithm is proposed that can be used to separate mixed signals of cognitiverecognition networks. In this algorithm, received mixed signal from cognition is firstcompressed in the algorithm, and then the received mixed signal is through edgedetection using wavelet edge detection method to obtain edge frequency. Next,bandpass filters are constructed according to the frequency edge. After that, blind sourceseparation method is applied to separate the mixed signal. The key points of thealgorithm are compressed cognition technology, wavelet edge detection technology andblind source separation technology. Simulation results show that this signal separationmethod can separate distinctly the main signal and disturbing signal and also has a good performance and low complexity.Fourthly, targeting at the difficulty of deciding mixed signals for secondary users innon-cooperative mode, a signal modulation method based on recognition modulation isproposed. Simulation results show that this algorithm could recognize three of fourkinds of source signals in the environment of analog multiple modulation mixed signals,and has fulfilled the tasks of division and recognition of non-cooperative mixed signalsin GSM, CDMA, and LTE. The recognition performance is good.Above all, the task of cognitive technologies on wireless spectrum environment is notonly detecting and deciding whether there is a signal or not in specific space andspecific time, but also various works such as signal separation and modes recognition.Aiming at these tasks, this paper explores wireless spectrum environment cognitiontechnologies with cognition convenience and efficiency.