Wideband Spectrum Sensing By Multi-step Sample Autocorrelation Detection In Cognitive Radio

The current spectrum management mainly adopts static allocation,which results in low spectrum utilization.Furthermore,With the rapid increasing demand on wireless communication services,the spectrum resource becomes scarcer in supply.To resolve the issues of low spectrum utilization and spectrum scarcity,It is an effective way to solve the problems by using the technology of software radio and cognitive radio(CR)to detect the idle spectrum for dynamic spectrum access.And,The spectrum detection technology is the key technology in CR,this thesis mainly researches this technology.Energy detection method is a simple signal source detection algorithm that firstly proposed by Urkowitz.However,the performance of energy detection is quite vulnerable to noise uncertainty.To overcome the shortcoming of Energy detection,Zeng and Liang proposes a covariance absolute value(called CAV)algorithm,the diagonal and off-diagonal elements in the statistical covariances matrix of the received signal are compared to determine the presence of primary users(called PU)without the knowledge of the noise power,the detection is robust to the noise uncertainty.When the working frequency bandwidth of the system is much larger than the maximum detection bandwidth of the device,it can not complete the full spectrum detection once.Wideband spectrum sensing capability can be enabled in software radio platform(like GNU Radio)by using multi-step frequency domain energy detection.But it is very sensitive to the noise power uncertainty.Especially in the low signal-noise-ratio(SNR),the detection performance is significantly decreased.Considering the signal sample autocorrelation detection is robust to noise uncertainty,to improve the detection performance of Cognitive Radio(called CR),we propose a step-by-step wideband spectrum sensing method based on signal sample autocorrelation.The structures of the thesis are organized as follows:The background and research significance of my paper describes in the first chapter,the existing problems and its solutions in the field of spectrum detection are introduced firstly.Then overview the cognitive radio technology and the research status of spectrum sensing detection technology.we describes the cognitive radio spectrum detection technology and classification them in the second chapter,And some main spectrum sensing detection is briefly analyzed.Then several classical spectrum sensing detections are compared.Finally,the impact of noise power uncertainty on the Energy detection is briefly introduced.In the third chapter,the basic principles of signal detection are given,Then we introduce the signal sample autocorrelation spectrum sensing algorithm,and obtain the theoretical formula of the decision threshold.Finally,a new method to determine the detection threshold based on detection probability and false alarm probability is proposed.To improve the detection performance in software radio platform(like GNU Radio),in the fourth chapter,we propose a novel approach of wideband spectrum sensing by multi-step sample autocorrelation detection.In addition,to balance the bandwidth resolution and the detection speed,we further propose to apply variable step values in two stages of wideband spectrum sensing based on signal sample autocorrelation detection,which can obtain higher detection resolution of the frequency bandwidth,and can shorten the detection time.Then,we simulate whole procedure for wideband spectrum sensing and the related detection performance of the algorithm using MATLAB.