Research Of Spectrum Sensing Technology Based On Noise Uncertainty And Sensing Strategy Optimization

The time-varying characteristics and non-unification of noise result in uncertainty of its numerical characteristics,namely,noise uncertainty.Noise uncertainty influences accuracy and stability of idle spectrum sensing for target spectrum band.The more complicated the environment is,the higher the mistake rate and false alarm rate of target spectrum band get.Noise uncertainty in wireless communication has been a development bottleneck of spectrum sensing technology.To clear the influence of noise uncertainty on cognitive radio as far as possible and improve accuracy and stability of spectrum sensing,several noise uncertainty models applicable to wireless communication are proposed in this paper,where the influence of noise uncertainty on spectrum sensing is restrained by reasonable modeling for background noise.This greatly improves sensing accuracy and reliability.The sensing strategy of spectrum sensing considers the primary factors that influence sensing performance and they determine the majorization direction and upper limit of sensing performance.Most alternative sensing strategies of spectrum sensing are local and only consider part of the main elements that influence sensing performance.This lowers the adaptability and robustness of cognitive system,which goes against the application and popularization of spectrum sensing in complex networks.As a result,research on global sensing strategy deserves the consideration of the whole primary factor influencing sensing performance.It helps to further perfect spectrum sensing theory,promote standardization of spectrum sensing and provides reference for spectrum sensing optimization under different circumstances.Motived by these,this paper is devoted to spectrum sensing research based on noise uncertainty and sensing strategy optimization.In this paper,sensing performance is improved by reducing noise contribution of received signal and optimizing current sensing strategy.This will provide reference to optimize sensing performance in various scenarios.The main work and contribution of this paper is concluded as follows:1.Low-rank noise model:we model spectrum of received signal as low-rank and sparse component,where noise spectrum is low-rank and signal spectrum is sparse.By removing low-rank part of received signal spectrum,the influence of noise uncertainty on spectrum sensing is suppressed and sensing accuracy is improved in return.2.Synergy noise model based on correlation and variance:we design two different adjustment factors and respectively add them to correlation factor and variance factor of observed signal.Then the product of signal correlation and its variance is time-invariant,and so is the product of noise correlation and its variance.As a result,the influence from noise uncertainty on sensing performance is indirectly controlled.3.Subspace noise model:we decompose received signal into two orthogonal subspaces,signal subspace and noise subspace.Noise subspace only contains noise in the environment of wireless communication while signal subspace contains the whole signal as well as little noise.By removing noise subspace,the residual signal subspace is used to detect idle spectrum band with a higher sensing accuracy and stability.4.Spectrum sensing based on noise uncertainty in multiple user co-existence condition:on the basis of subspace noise model,we propose a spectrum sensing scheme considering multiple user co-existence and analyze its detection probability and false alarm probability in theory.5.Multiple-stage spectrum sensing based on noise uncertainty:we explore two-stage blind spectrum sensing with energy detector and covariance matrix detector based on effective evaluation of environmental signal-to-noise ratio(SNR),where covariance matrix detector is applied to spectrum sensing for the suppression of noise uncertainty in strong noise conditions.6.Spectrum sensing based on sensing strategy optimization:we explore single node based global sensing strategy considering detection probability,false alarm probability,sensing time and system throughput.In addition,single node based global sensing strategy is discussed with sensing performance,sensing time,system throughput and computational complexity.