一种改进的延迟型LMS自适应滤波器及其FPGA实现

摘要

LMS(最小均方误差算法)自适应滤波器结构简单、运算量小,在地质勘探信号处理等实际中有着广泛的应用.针对普通结构的LMS自适应滤波器运行速度相对较慢的缺点,采用割集时序重整技术和流水线松弛等技术,对普通结构的LMS自适应滤波器进行了优化,得到了一种新的延迟型结构LMS自适应滤波器.利用DSP Builder工具建立了4阶普通结构LMS自适应滤波器和改进的延迟结构LMS自适应滤波器模型并进行了一系列的仿真与对比分析.结果表明:改进的延迟型LMS自适应滤波器速度更快,耗费逻辑资源少.%Least Mean Square adaptive filter has been widely used such as geological exploration signal processing etc. because of its simple structure and small calculation amount. Aiming at the shortcomings of slow speed of the ordinary structure LMS adaptive filter, using the cutset timing reforming technology and pipeline relaxation , the structure of the LMS adaptive filter was optimized, an improved delayed structure of the LMS adaptive filter was acquired at first in the paper. Then an model of four - order ordinary and improved delayed LMS structure adaptive filter was built based on the DSP Builder tool and a series of simulation and comparative analysis was fulfilled with a variety of EDA tools respectively. The results show that the improved delayed structure LMS adaptive filter has a faster response speed than the ordinary structure, and the resources were used little less.