The Research Of Sparse Channel Estimation Algorithm Based On Compressed Sensing

High-rate data communication over a multipath wireless channel often requiresthat the channel response be known at the receiver, so learning the disseminationcharacters of wireless channel and estimating the parameters are very important todigital wireless communication system. The current channel estimation methods aremostly based on training sequence（pilot）, which probe the channel with signalsknown in time or frequency, and reconstruct the channel response from thecorresponding output signals. These methods may bring a problem: the pilot occupiedbandwidth, so the spectral efficiency of the system would be decreased. How can weuse fewer pilots to estimate the channel which does not affect channel estimationperformance in the premise is an issue that deserves research. However, physicalarguments and experimental evidence suggest that wireless channels encountered inpractice exhibit a sparse multipath structure that gets pronounced as in high-rate datacommunication system. Actually, the traditional channel estimation algorithm doen’tconsider the channel’s inherent sparse characteristic, then, how to dig the channelfully and adopt more effective way to estimate sparse channel attracted theresearchers’ attention.In recent years, the concept of compressed sensing(CS) has been put forward, itdemonstrats that if a signal is sparse (or nearly sparse) in some domain, then we canrecover this high-dimensional but sparse vectors from relatively few samples, whichcan be accomplished by solving a tractable convex optimization in the sparsedecomposition theory. Compressed sensing has recently emerged as a powerful signalacquisition paradigm. Besides, sparse wireless channels meet the premise of thistheory, so compressed sensing also brings us a new method for sparse channelestimation, which has a good application prospect.Follow this line, the thesis focuses the research on the channel estimation basedon the compressed sensing. The main work is as follows:Firstly, the thesis analyzed the primary characters of wireless channels.Secondly, the thesis introduced the theoretical framework of compressed sensingas well as its reconfiguration algorithm carefully.Thirdly, the thesis combined channel estimation problem and the sparse signalreconstruction in compressed sensing together, then we introduced some sparse channel estimation methods based on basis pursuit algorithm and orthogonal matchingpursuit algorithm, we also compared these methods with the traditional channelestimation methods by simulation, the results verified the former has better estimationperformance in sparse channel estimation.Finally, a new estimation method based on compressive sampling matchingpursuit algorithm for single-antenna frequency selective sparse channel was proposedin this thesis. The results of simulation showed that this method achieved a targetreconstruction error by shorter training sequence (fewer pilots) than orthogonalmatching pursuit algorithm and least square error algorithm, which had higherspectrum efficiency and estimation accuracy.