Research On Spectrum Sensing Based On Data Fusion And Compressed Sensing Technology

With the rapid development of communication industry, the spectrum resources becomesincreasingly valuable now，so cognitive radio technologies have emerged to address thisproblem, the technology accesses licensed bands for communication in an "opportunistic"way, it can be a good solution to the lack of spectrum resources and is able to improve theutilization of spectrum resources effectively. Before the use of idle spectrum, cognitive usersneed to dynamically search for a free band, thus spectrum detection technology is the coretechnology of cognitive radio, high reliability and high-quality spectrum detection pattern is aprerequisite of cognitive network to work properly.This paper first introduces the research background and current status of cognitive radiotechnology, analysis different spectrum detection technology models and their merit anddemerit, and combine compressed sensing technology and spectrum detection technology insignal processing technology to solve the problem that wideband channel sampling hardwareproblem is difficult to achieve, we propose the spectrum detection technology based oncompressed sensing technology and introduce the research background and theoreticalfoundation of compressed sensing technology in details, and make a groundwork for theresearch work below.This paper we deduce and simulate the local spectrum detection theory of singlecognitive user, multiple cooperative spectrum detection scheme based on dual-thresholdaiming at the shortage of a single node detection is proposed, the program optimized for thetraditional dual-threshold spectrum detection, adopted double threshold pattern of multipledetection and improved the structure of collaboration cognitive network. The partition fusionrule is taken in the cluster head and "or" guidelines is used in the central node. At last, theproposed program was compared to the traditional cooperative detection scheme by thesimulation, verifying the improvement of its detection performance.In addition, this paper analyzes two commonly used compressed sensing reconstructionalgorithm: the algorithm based on greedy iterative algorithm and based on Bayesian statistics.First of all, according to the commonly used MP, OMP, StOMP reconstruction algorithm, weproposed a spectrum detection algorithm based on these three algorithms. Then we analyzethe reason of MP detection algorithm eigenvectors volatile and provide an improved scheme. The simulation experiments explained the impact of different parameters on thereconstruction performance and detection performance, and verified that the improvedscheme can effectively compress the wideband channel and detect spectral. Secondly, weanalyzed the Bayesian compressed sensing spectrum detection technology effect of differentparameters based on the method of counting zero, and bring in the observation matrixadaptive algorithm through the process to build measurement matrix, forming the optimizedBayesian compressed sensing spectrum detection scheme. At last, we analysis the impact ofdifferent parameters on the detection performance through the simulation experiments, andthe proposed program was compared to the traditional Bayesian compressed sensing schemeby the simulation, verifying the improvement of its detection performance.