摘要
针对高性能二维线性相位有限脉冲响应(finite impulse response,FIR)滤波器设计,随着阶数增加导致计算和存储压力过大的问题,提出一种稀疏二维FIR滤波器设计的迭代方法。稀疏二维FIR滤波器设计问题被转化为神经网络的李雅普诺夫能量函数最小化问题。在每次迭代中,利用加权Frobenius范数的正交匹配追踪(orthogonal matching pursuit,OMP)算法寻找滤波器中非零系数的位置,使用Hopfield神经网络计算非零系数。仿真结果表明,该方法能够有效增加二维FIR滤波器的稀疏度,稀疏效果优于已有的稀疏滤波器设计算法。
As to the high performance two-dimensional linear phase finite impulse response(FIR)filters design,the calculation and storage pressure is too large due to the order increase.An iterative method for sparse two-dimensional(2-D)FIR filters design was then proposed.The sparse 2-D FIR filter design problem was transformed into the Lyapunov energy function minimization problem of neural networks.In each iteration,positions of the non-zero coefficients of a filter were found using the orthogonal matching pursuit(OMP)algorithm with weighted Frobenius norm,and the non-zero coefficients were computed using the Hopfield neural network.Simulation results show that the proposed method can effectively increase the sparsity of 2-D FIR filter,and the sparsity is higher than that of the existing sparse filter design algorithms.
作者
李怡
赵加祥
徐微
LI Yi;ZHAO Jia-xiang;XU Wei(College of Electronic Information and Optical Engineering,Nankai University,Tianjin 300381,China;School of Electronics and Information Engineering,Tiangong University,Tianjin 300387,China;Tianjin Key Laboratory of Optoelectronic Detection Technology and System,Tiangong University,Tianjin 300387,China)
出处
《计算机工程与设计》
北大核心
2025年第7期1841-1847,共7页
Computer Engineering and Design
基金
国家自然科学基金项目(61671254)。
作者简介
李怡(1992-),女,浙江杭州人,博士研究生,研究方向为基于深度学习的稀疏FIR滤波器设计;通讯作者:赵加祥(1961-),男,福建福州人,博士,教授,研究方向为微移动系统的控制算法、信道编码算法;徐微(1984-),女,河北保定人,博士,副教授,研究方向为多信号处理、数字滤波器设计。E-mail:zhaojx@nankai.edu.cn。
