现代数字信号处理-基于LMS线性预测剖析.doc

现代信号处理 基于LMS 算法的线性预测估计 小组组长： 刘鑫 (150408520845) 小组成员： 刘芳(150408520846) 万娇150408520849) 郑熔(150408520848) 郭俊(150408520852) 任课教师： 聂文滨 教师所在学院： 信息工程学院 时间：现代电子技术发展日新月异，随着数字时代的到来，各种各样的产品层出不穷，在追求高性能的同时，高效的算法越来越得到人们的青睐本文对其改进算法进行了研究主要有的线性预测，（），，。其中，线性预测是根据已有采样点按照函数计算未来某一的数学方法，线性预测可分为前向性预测和后向性预测在线性预测中维纳滤波应用很广泛，包括线性预测器原理，线性预测与AR模型的关系以及线性预测器的AR模型功率谱估计。在系统分析中，线性预测可以看作是建模或者最优化的一部分另外，是不同的根据参数模型功率谱估计的思想， 最小均方误差准则线性预测系数或LPC系数从而进行线性预测分析。预测软件仿真Abstract Development of modern electronic technology with each passing day, a variety of products emerge in endlessly with the arrival of digital age, in the pursuit of high performance at the same time, the efficient algorithm is more and more get the favour of people. This paper studies the Least Mean Square (LMS) algorithm and its improved algorithm , mainly include the linear prediction of LMS algorithm, the decorrelation LMS algorithm (including the temporal decorrelation algorithm and transform domain decorrelation algorithm), adaptive LMS algorithm, and using matlab software to simulat several algorithms, through the simulation results show the performance of LMS algorithm with all kinds of intuitive. The linear prediction is a mathematical method to calculate a future discrete signal according to the sampling points by the linear function ,linear prediction can be divided into forward prediction and backward prediction. Wiener filtering is widely used in the linear prediction ,including the principle of linear predictor, the relationship between linear prediction and auto-regressive(AR) model and AR model power spectrum estimation of the linear predictor. In system analysis, linear prediction can be seen as part of the mathematical modeling and optimization. In addition, the different results when the LMS algorithm is adopted to affect the LMS algorithm is different. According to the ideas of the parameter model of power spectrum estimation, using the LMS