基于格型结构LMS算法的自适应预测

The Adaptive Prediction Based on Lattice Structure LMS Algorithm

自适应预测在语言和图像编码中有着大量的应用。本文重点研究LMS算法的一步线性预测。其间推导了LMS算法在格型结构中的表达式,并验证了二阶格型结构的LMS算法收敛过程,也在格型结构系统中验证了LMS算法在一步前向预测上的准确性。结果表明,在格型结构中使用LMS算法进行一步前向预测,能够达到较高的准确性。

The adaptive prediction has a lot of applications in language and image coding.This paper focused on the one-step linear prediction of LMS algorithm.In the meantime,the expression of the LMS algorithm in the lattice structure was derived,and the convergence process of the LMS algorithm of the second-order lattice structure was verified,and the accuracy of the LMS algorithm in one-step forward prediction was also verified in the lattice structure system.The above verification shows that using the LMS algorithm for one-step forward prediction in the lattice structure can achieve higher accuracy.