平稳环境下LMS算法优化分析及其FPGA实现  

Optimization Analysis of LMS Algorithm in Stationary Environment and Its FPGA Implementation

作　　者：高金定[1,2] 

机构地区：[1]中南大学地球科学与信息物理学院,湖南长沙410083 [2]湖南涉外经济学院电气与信息工程学院,湖南长沙410205

出　　处：《压电与声光》2013年第3期445-447,451,共4页Piezoelectrics & Acoustooptics

基　　金：湖南省教育厅优秀青年基金资助项目(09B058);湖南省教育厅重点基金资助项目(10A068)

摘　　要：高速自适应滤波器的设计及其现场可编程门阵列(FPGA)实现一直是信号处理领域研究的热点。普通最小均方误差算法(LMS)自适应滤波器运行速度受到计算滤波输出和系数更新时间限制,制约了其信号处理的速度。对于平稳环境,通过对自适应滤波器系数更新方程进行前瞻和松弛近似,对LMS自适应算法进行了优化分析,得到了一种改进的LMS自适应滤波器结构。利用DSP Builder工具建立了四阶改进结构的LMS自适应滤波器模型并进行了一系列的仿真,结合多种电子设计自动化(EDA)工具,最终在EP2C35型FPGA上得到了最高响应速度60.07MHz的高速自适应滤波器。结果表明,改进的LMS自适应滤波器速度较一般结构滤波器快,但耗费了较多逻辑资源。The design and FPGA implementation of high-speed adaptive filter has been a hot research field of signal processing. However, the running speed of the ordinary LMS adaptive filter was limited by the calculating fil- ter output and the coefficients updating time. For stationary environment, using prospective technique and relaxation approximation, adaptive filter coefficient update equations was optimized, an improved structure of the LMS adap- tive filter was acquired at first in this paper. Then an model of four-order improved LMS structure adaptive filter was built based on the DSP Builder tool and a series of simulation was fulfilled with a variety of EDA tools, ulti- mately, an high-speed adaptive filter with 60.07 MHz was implemented on the EP2C35 type FPGA. The results show that the improved structure LMS adaptive filter has a faster response speed than the general structure, but the resources were used more.

关 键 词：平稳环境 最小均方误差算法(LMS) 自适应滤波器 现场可编程门阵列(FPGA) 前瞻技术 松弛近似 仿真 

分 类 号：TN911.72[电子电信—通信与信息系统] TP319[电子电信—信息与通信工程]