Channel Eastimation Of Cognitive Radio Based On Modified Compressive Sensing Matching Pursuit Algorithm

With the rapid growth of wireless communication, the demand of people for wirelessspectrum resources is becoming greater and greater increasingly.The cognitive radio hasbeen widely concerned as an effective technology to resolve the contradiction betweenlimited spectrum resources and spectrum demand.And the performance indexes of thewireless channel estimation method which is regarded as the key technology of cognitiveradio relates directly to the communication of cognitive radio.In recent years,compressive sensing has became the research focus in the world ofsignal processing and wireless communication.The time domain model of mutipathchannel can be equivalent to a transversal filter of which tap is sparse distributed,and itshows that the feasibility of channel estimation by means of compressive sensing.In thispaper,we target reducing system overhead and enhancing performance of channelestimation,study the channel estimation algorithm of cognitive radio based on compressivesensing.Firstly, analysis and summary the existing channel estimation methods on the basis ofthe analysis of cognitive radio and compressive sensing.Then discuss feasibility ofchannel estimation for cognitive radio with compressive sensing.Secondly,replace the compressive sensing matching pursuit algorithm with the weakmatching pursuit algorithm to thansform the optimization problem into suboptimal.Andfurther at the stage of the searching of optimal atom,combine the global optimizationcharacteristics of PSO,come up with a optimizing strategy for the weak matching pursuitalgorithm which is a adaptive simplified particle swarm based on correction factor,whichalgorithm regards the average size of seed spacing between particles as the triggeringcondition,and then adaptively adjust intertia weight、the average individual extremum andglobal extremum.At last,the algorithm can make particles find the global optimal solutionquickly and accurately.The simulation results vertifiy the effectiveness and the rapidity ofthis modified algorithm.Finally,by means of choosing the wireless transmission channel which under theNC-OFDM(Non-Contiguous Orthogonal Frequency Division Multiplexing) system as research background,we proposed a new algorithm that is intrduced two new mechanismcalled fast selecting and survival of the fittest.And discuss the effect of the number of thepilot frequency to the channel estimation performance.We have simulated under thecondition of broadband and the narrowband interference.The results indicate that the newalgorithm,compare to LS and the classical orthogonal matching pursuit algorithm,canreconstruct the channel preferably,and get smaller MSE and lower BER,besides it onlyspend fewer pilot frequency to get the same channel estimation performance.