论文部分内容阅读
自适应滤波器为一个未知系统的建模提供了一种简单,实用的方法。可以在不知道系统的理论模型的条件下,通过测量和学习,实现对未知系统的最佳拟和。本文简要介绍了自适应滤波在信号处理系统建模中的应用,着重讨论了IIR递推结构自适应滤波器的LMS算法。在此理论基础上阐述了一种改进的LMS算法。新算法利用误差信号的相关值调节算法步长,解决了收敛时间和稳态误差的矛盾,并且不受已经存在的不相关噪声的干扰。
Adaptive filters provide a simple and practical way to model an unknown system. The best fit of unknown system can be achieved by measuring and learning without knowing the theoretical model of the system. This article briefly introduces the application of adaptive filtering in signal processing system modeling, and focuses on the LMS algorithm of IIR recursive structure adaptive filter. Based on this theory, an improved LMS algorithm is presented. The new algorithm uses the correlation value of the error signal to adjust the algorithm step size and solves the contradiction between convergence time and steady-state error, and is not disturbed by the existing uncorrelated noise.