改进的变步长LMS改进算法  

Improved variable step size LMS improved algorithm

作　　者：张晶晶[1] 周菲菲[1] 许帅[1] 

机构地区：[1]装甲兵工程学院,北京100072

出　　处：《现代电子技术》2014年第1期11-13,共3页Modern Electronics Technique

摘　　要：在基于Lorentzian函数的变步长LMS自适应滤波算法的基础上,进行进一步改进,提出了一种新的自适应LMS滤波算法,通过建立新的误差信号e(n)与变步长因子μ(n)之间的关系,消除不相关噪声的影响。并用Matlab对其进行仿真验证,表明该算法解决了收敛速度和稳态误差之间的矛盾,在保证算法的计算复杂度较低的同时,使得算法的抗干扰能力进一步提高,适用于低信噪比条件下的信号提取及滤波,为实际应用提供了更大的灵活性。A new adaptive LMS filtering algorithm is proposed by improving the Lorentzian function-based variable step size LMS adaptive filtering algorithm. Through establishing the relationship between error signal e（n） and variable step size factor μ（n） to undo the uncorrelated noise. The algorithm is simulated by Matlab, and the results show that＇the improved algorithm has solved the contradiction between the convergence rate and steady-state error. The anti-interference ability of the algorithm is improved while guaranteeing the computational complexity is lower. It is suitable for low SNR signal extraction and filtering, and provides greater flexibility for the practical application.

关 键 词：自适应滤波 噪声抵消 收敛速度 抗干扰能力 

分 类 号：TN913.4[电子电信—通信与信息系统]