資源描述:
《基于基追蹤_Moore_Penrose逆矩陣算法的稀疏信號(hào)重構(gòu)》由會(huì)員上傳分享,免費(fèi)在線閱讀,更多相關(guān)內(nèi)容在行業(yè)資料-天天文庫(kù)。
1、第35卷第2期電子與信息學(xué)報(bào)Vol.35No.22013年2月JournalofElectronics&InformationTechnologyFeb.2013基于基追蹤-Moore-Penrose逆矩陣算法的稀疏信號(hào)重構(gòu)*張曉偉李明左磊(西安電子科技大學(xué)雷達(dá)信號(hào)處理國(guó)家重點(diǎn)實(shí)驗(yàn)室西安710071)摘要:壓縮感知(CompressedSensing,CS)稀疏信號(hào)重構(gòu)其本質(zhì)就是在稀疏約束條件下求解欠定線性方程組,基于迭代加權(quán)Lp-<£=(0p1,p2)類范數(shù)算法減小重構(gòu)誤差成為近來稀疏信號(hào)重構(gòu)熱點(diǎn)之一。該文提出了基追蹤-Moore-Penrose逆矩陣(BasisPurs
2、uit-Moore-PenroseInverseMatrix,BP-MPIM)算法:(1)由基追蹤(BasisPursuit,BP)算法得到稀疏信號(hào)非零元素位置(亦稱支撐集,對(duì)應(yīng)于測(cè)量矩陣的列);(2)通過求解由支撐集所對(duì)應(yīng)測(cè)量矩陣的子矩陣和CS測(cè)量值組成的超定線性方程組實(shí)現(xiàn)稀疏信號(hào)重構(gòu),并證明了由此重構(gòu)的稀疏信號(hào)是其唯一最小二次范數(shù)解。仿真的稀疏信號(hào)和實(shí)測(cè)寬帶雷達(dá)回波信號(hào)脈沖壓縮結(jié)果表明,和原來算法相比,新算法具有更小的重構(gòu)誤差,且誤差只存在于其支撐集內(nèi)。關(guān)鍵詞:雷達(dá)信號(hào)處理;壓縮感知(CS);信號(hào)重構(gòu);基追蹤(BP);Moore-Penrose逆矩陣中圖分類號(hào):TN95
3、7.51文獻(xiàn)標(biāo)識(shí)碼:A文章編號(hào):1009-5896(2013)02-0388-06DOI:10.3724/SP.J.1146.2012.00238SparseSignalReconstructionBasedonBasisPursuit-Moore-PenroseInverseMatrixZhangXiao-weiLiMingZuoLei(NationalLaboratoryofRadarSignalProcessing,XidianUniversity,Xi’an710071,China)Abstract:Thesparsesignalreconstructionwith
4、CompressedSensing(CS)isactuallysolvingasystemofunderdeterminedlinearequationswithinthesignalsparsity,ofwhichonefocusistoreducerecoveryerrorsbythetypeofiterativelyweightedLp-<£=(0pp1,2)algorithmsrecently.TheBasisPursuit-Moore-PenroseInverseMatrix(BP-MPIM)algorithmisproposedinthispaper.First
5、,nonzeroelementcoordinatesofthesparsesignalareacquiredbythebasispursuitalgorithm,whicharerenamedwiththesparsesignalsupportset(correspondingwithcolumnsofthemeasurematrix).Then,thesparsesignalrecoveryissolvedfromasetofsuperdeterminedlinearequations,whichiscomposedofthesubmatrixofthesamplingm
6、atrixandcompressedsensingmeasurements.Atthesametime,itisprovedthatthereconstructionofsparsesignalsbythisnewalgorithmistheoneandonlyminimizeL-2norm.Bothsimulativesparsesignalsandpulsecompresseddataofwidebandradarechoesindicatethatthenewalgorithmhaslessrecoveryerrorsthanthepreviousalgorithms
7、,whicharejustinthesupportset.Keywords:Radarsignalprocessing;CompressedSensing(CS);Signalreconstruction;BasisPursuit(BP);Moore-PenroseInverseMatrix(MPIM)?1引言知,回波信號(hào)為發(fā)射信號(hào)和具有有限個(gè)散射點(diǎn)目標(biāo)的響應(yīng)函數(shù)卷積,它經(jīng)過匹配濾波后表現(xiàn)為1維距通過各種各樣的傳感器,人們可以從自然界中離像,其脈沖壓縮結(jié)果中較大值是有限的。隨著發(fā)獲得不同的信號(hào),但對(duì)于獲得的相當(dāng)多信