Research Of Spectrum Sensing Techniques In Cognitive Radio

With the development of wireless technology and the appearance of newwireless equipments,the radio spectrum resource is getting more and morestrained.The current spectrum allocation scheme uses the static spectrumallocation,which causes low spectrum efficiency.Cognitive Radio accesses thespectrum in some “opportunistic way”， which does not affect the normalcommunication of the authorized frequency band,and uses the spectrumdynamically,improving the spectrum efficiency.Spectrum sensing is the premise ofrealizing cognitive radio.It keeps monitoring the channel state,detects the spectrumholes,and vacates the occupied spectrum band when primary user signal appearsagain to cause no interference.The traditional spectrum sensing employs energy detection algorithm.Thealgorithm is a noncoherent detection and is easy to be realized.But its detectionperformance falls sharply in low SNR,and is affected by the uncertainty of noisepower.In this paper we utilize the spectrum sensing algorithm based on samplecovariance matrix,which overcomes the disadvantages above,and improves thedetection performance.In recent years,RMT(Random Matrix Theory)has been used in signaldetection theory widely.We adopt the RMT and get the sample covariance matrix ofthe received signals,then we use the ratio between the maximum eigenvalue of thematrix and the average power of the received signals and the ratio between themaximum and the minimum eigenvalue as the test statistics respectively to realizespectrum sensing.We compare the algorithm with the traditional cooperativespectrum sensing based on energy detection through the experimentalsimulation.Then we know the algorithms increase the detection probabilities withno substantial improvement in computational complexity.Finally,contraposing the problem of treating the cognitive usersequally,without considering the differences of detection reliabilities in traditionalspectrum sensing,we employ the weighted-collaborative spectrum sensing and thesensing algorithm based on D-S theory,the algorithms overcome the shortcomingsof the traditional spectrum sensing and improve the detection probability.