基于FPGA的自适应噪声消除DLMS算法

摘要

研究了自适应噪声消除DLMS(Delay Least Mean Square)算法在实现时的速度和运算复杂度问题,提出了二进制树直接结构实现DLMS算法.算法采用了割集重定时技术和流水线结构,对延迟模块重新分割,使系统关键路径降到最低的同时具有较快的收敛速度.仿真结果表明与原有结构相比,新算法结构提高运算速度近3倍,较好地消除了含噪信号中的噪声.%This paper studied the problems of speed and computational complexity which the DLMS( delay least mean square) algorithm in realization for adaptive noise cancellation, presented the binary tree direct form realization of DLMS algorithm. The algorithm adopted cut-set retimed techniques and pipeline structure,repartition the delay modules,made the system critical path to a minimum and at the same time with higher convergence rate. The simulation results show that compared with the original structure,the structure of the new algorithm to improve operation speed nearly three times,can eliminate the noise in the signal containing noise.