Research On Non-reconstruction Spectrum Sensing Algorithms Based On Compressive Sensing

Wireless spectrum,as a limited and non-renewable natural resource,can not fully meet the needs of wireless communication and the scarcity of wireless spectrum resources has become an urgent problem to be solved with the rapid development of wireless communication and the rapid increase of high-speed data transmission requirements.However,many spectrum resources allocated statically are underutilized,resulting in unnecessary waste and unbalanced use of spectrum resources.Cognitive radio(CR)allows secondary user(SU)who is not allocated spectrum resources to utilize the idle spectrum of primary user(PU),which can make full use of spectrum resources.In CR,spectrum sensing is the first step,and it is necessary to detect spectral holes that can be used by the SU without interfering with the PU.The traditional idea of using compressed sensing for spectrum sensing is to compress the signal and then reconstruct the signal for spectrum detection.According to CS theory,signal reconstruction is the most computational part of compressed sensing consumption.It is important to study spectrum sensing under non-reconstruction framework to reduce the computational complexity of spectrum sensing algorithm.A non-reconstruction spectrum sensing algorithm for single node based on compressive sensing(CS)is proposed to reduce the computational complexity under reconstruction and the influence of spectrum reconstruction accuracy on the decision result.In single node spectrum sensing(SSS)algorithm,the channels are first divided into channel groups to reduce the computational complexity of the sampled covariance matrix(SCM).Then the diagonal values of the SCM are obtained,and detection performance is further improved by a de-noising algorithm.The simulation results show that the SSS can effectively detect the spectrum holes.SSS may face the challenges that detection performance is degraded if there are deep fading channels and hidden terminals.Cooperative multiple nodes spectrum sensing(CMSS)is proposed to improve detection performance.Nonreconstruction CMSS can reduce the computational complexity of the SCM and improve the detection performance by dividing the channel into channel groups to obtain the value of the diagonal of the fused SCM.The simulation results show that CMSS can effectively detect the spectrum holes and the CMSS detection accuracy is higher.A two-stage spectrum sensing algorithm based on reputation hierarchy is proposed to solve the problem of spectrum sensing data falsification(SSDF)attack in CMSS.The sensing data(SD)of the secondary users and the SD of the honestly secondary user(HSU)are verified,then the reputation is leveled and the weight of the user in the fusion center(FC)is updated to reduce the influence of the detection performance degradation caused by the SSDF attack.The simulation results show that based on the nonreconstruction CMSS algorithm,the two-stage spectrum sensing algorithm based on the reputation hierarchy can improve effectively the detection performance when there is SSDF attack.