电信科1002班田苗苗

《电信科1002班田苗苗》由会员上传分享，免费在线阅读，更多相关内容在应用文档-天天文库。

1、11题目:数字图像处理一.摘要翻译与分析:Abstractx1----Anewnoniterativealgorithmisproposedtoestimatethesampledimpulseresponseofunknownchannels.Withtheproperchoiceofthetrainingsequence（训练序列）,implementationoftheproposedchannelestimator（估计器）requiresonlyadditionsandsubtractions,i.e.,nomu

2、ltiplicationsordivisionsareneeded.Moreover,thestructureoftheestimatorissosimplethatitcanbeeasilyimplementedusinganordinarymicroprocessorwithminimalstorage.Thechannelestimatecanserveinreducingtheeffectsofintersymbolinterferenceeitherbythemaximumlikelihoodsequencees

3、timator(MLSE最大似然序列估计器)usingtheViterbialgorithm（维特比算法）,orbychannelequalizationwithadirectsolutiontoobtaintheoptimumequalizertaps（抽头）.AnewprocedureforthelattercaseisproposedhereusingtheLevinson-Trenchalgorithmforfaststart-up（启动）ofadaptiveequalizersinnoisyenvironment

4、s.Theperformanceoftheproposedalgorithmisevaluatedthroughsimulationanditiscomparedtosomeoftheexistingtechniques.Computationalefficiencyisalsotakenintoaccount.Resultsofsimulationshowthesuperiority（优越性）oftheproposedscheme.摘要翻译：一种新的非迭代算法的提出是为了计算采样脉冲响应的未知的信道。选择适当的训练序列，

5、信道估计的实现只需要加法和乘法，而不需要其他运算器。此外，估计器的结构非常简单就可以使用普通的微处理器的最小存储容易实现。信道估计可以通过基于维特比算法的最大似然序列估计器，或者用以获得最佳的均衡器抽头直接求解均衡信道。对于后者的情况下提出了一个新的程序使用莱文森算法快速启动在嘈杂的环境中的自适应均衡器。所提出的算法的性能评价通过仿真和比较了一些现有的技术计算效率也考虑。仿真结果表明所提出的方案的优越性。摘要要素分析：目的：计算未知信道的脉冲响应方法：非迭代算法结果：成本低，易实现结论：对专业词汇翻译准确优点点评：Abst

6、ractx2—Thedevelopmentofanadaptiveinfiniteimpulseresponse(IIR无限脉冲响应)linearequalizerisdescribed.UsingdiscretetimeWienerfilteringtheory,aclosedform（封闭式）fortheoptimummean-squareerrorIIRfilterisderived.Aperformancecomparisonusingbothminimumandnon-minimumphasechannelsin

7、dicatesthecomplexity/performanceadvantagesinherentintheIIRsystemcomparedtoanoptimumfiniteimpulseresponse(FIR有限脉冲响应)solution.Theminimumphasespectralfactorization（谱分解）,whichisanintegralpartofthederivationoftheIIRequalizer,maybecircumvented（设法克服，避免）throughtheuseofaKa

8、lmanequalizersuchasthatoriginallyproposedbyLawrenceandKaufman.ThestructureismadeadaptivebyusingasystemidentificationalgorithmoperatinginparallelwithaKal