LMS filtering algorithm based on linear regular transformation

• Main Authors: QIU WEI, SHI PEIFENG
• Format: Patent
• Language: Chinese; English
• Published: 29.06.2016
• Subjects: BASIC ELECTRONIC CIRCUITRY; ELECTRICITY; IMPEDANCE NETWORKS, e.g. RESONANT CIRCUITS; RESONATORS

The invention relates to an LMS filtering algorithm based on linear regular transformation, and the algorithm sequentially comprises the following steps: 1, calculating the LCT of an input signal x(n) and a target signal d(n) at an optimal parameter A; 2, calculating an error vector EA(k) of a time index k; 3, calculating an updated weight coefficient vector W(k) according to a recursion algorithm; 4, adding one to the time index k, and returning to step 2 for loop computing till the algorithm is stable; 5, carrying out the optimal parameter inverse LCT of an output signal after the algorithm is stable, and obtaining a time domain output signal after filtering. The algorithm is very low in requirements for the signal to noise ratio of a signal, and is superior in the detection and filtering of chirp signals. 本发明涉及一种基于LCT的LMS自适应滤波算法，依次包括以下步骤：步骤一、计算输入信号x(n)和目标信号d(n)在最佳参数A的LCT；步骤二、计算时间指数k的误差矢量E(k)；步骤三、按递归算法计算更新后的权系数矢量W(k)；步骤四、将时间指数k加1，返回步骤二循环计算，直到算法达到稳定；步骤五、对算法稳定后的输出信号进行最佳参数的逆LCT，得到滤波后的时域输出信号。本发明的算法对信号的信噪比要求很低，在