基于QPSO算法的准完全重构余弦调制滤波器组的优化设计  

Near-Perfect-Reconstruction Cosine Modulated Filter Banks Optimization Design Based on Quantum-Behaved Particle Swarm Optimization Algorithm

作　　者：谭方青[1] 张天骐[1] 高春霞[1] 黄烈超[1] 

机构地区：[1]重庆邮电大学信号与信息处理重点实验室,重庆400065

出　　处：《电信科学》2011年第7期80-85,共6页Telecommunications Science

基　　金：国家自然科学基金资助项目(No.61071196);国家自然科学基金-中物院NSAF联合基金资助项目(No.10776040);教育部新世纪优秀人才支持计划资助项目(No.NCET100927);信号与信息处理重庆市市级重点实验室建设项目(No.2009CA2003);重庆市自然科学基金资助项目(No.2009BB2287;No.2010BB2398;No.2010BB2411)

摘　　要：提出了一种准完全重构的余弦调制滤波器组的设计方法,使用优化方法设计原型滤波器。该方法固定原型滤波器的阻带截止频率为ωs=π/M,以通带截止频率ωp为参数变量,用量子粒子群优化算法(QPSO)优化满足重构条件的目标函数,间接设计原型滤波器,然后通过调制得到余弦调制滤波器组。稍微放宽余弦调制滤波器组的精确条件,从而大大降低了设计的复杂性,减少运行时间。仿真实验结果表明,该算法简单有效,可获得具有高阻带衰减、低混叠误差和重构误差的余弦调制滤波器组。In the paper,a new design method of near-perfect-reconstruction cosine modulated filter banks is proposed,the prototype filter is designed by optimization method which fixes the prototype filter stopband cutoff frequency,takes passband cutoff frequencyas a parameter variable,optimizes the target function which meets reconstruction condition,designs indirectly the prototype filter,and then obtains the cosine modulated filter banks by modulation.It appropriately relaxes the limit conditions of perfect reconstruction of CMFB,so obtains lower complexity and computing time.The simulation results demonstrate that the method proposed is simple and effective,can obtain high stopband attenuation,low aliasing distortion and reconstruction error banks.

关 键 词：准完全重构 余弦调制滤波器组 原型滤波器 量子粒子群优化算法 

分 类 号：TN713[电子电信—电路与系统]