摘要
为了解决分布式最小均方算法在输入信号相关性较高时收敛速度较慢、分布式仿射投影算法计算复杂度较高等问题,本文提出了两种分布式子带自适应滤波算法,即递增式和扩散式子带自适应滤波算法.分布式子带自适应滤波算法将节点信号进行子带分割来降低信号的相关性,从而加快收敛速度.由于用于子带分割的滤波器组中包含了抽取单元,所以分布式子带自适应滤波算法和对应的分布式最小均方算法的计算复杂度相近.仿真结果表明,与分布式最小均方算法相比,分布式子带自适应滤波算法具有更好的收敛性能.
To address the problem of slowconvergence rate of the distributed least mean square algorithms for highly correlated input signals and the problem of high computational complexity of the distributed affine projection algorithms,this paper proposes two distributed subband adaptive filtering algorithms,i. e.,the incremental and diffusion subband adaptive filtering algorithms. The distributed subband adaptive filtering algorithms partition the signals of nodes to reduce their correlation so that their convergence rate can be increased. The computational complexities of the distributed subband adaptive filtering algorithms are close to those of their corresponding distributed least mean square algorithms due to decimation operations included in the filter banks used for subband partition. Simulation results showthat the distributed subband adaptive filtering algorithms exhibit good performance as compared to the corresponding distributed least mean square algorithms.
引文
[1]Sayed A H,Tu S Y,et al.Diffusion strategies for adaptation and learning over netw orks[J].IEEE Signal Processing M agazine,2013,30(3):155-171.
[2]Lopes C G,Sayed A H.Incremental adaptive strategies over distributed netw orks[J].IEEE Transactions on Signal Processing,2007,55(8):4064-4077.
[3]Lopes C G,Sayed A H.Diffusion least-mean squares over adaptive netw orks:formulation and performance analysis[J].IEEE Transactions on Signal Processing,2008,56(7):3122-3136.
[4]Cattivelli F S,Sayed A H.Diffusion LMS strategies for distributed estimation[J].IEEE Transactions on Signal Processing,2010,58(3):1035-1048.
[5]Liu Y,Li C,Zhang Z.Diffusion sparse least-mean squares over netw orks[J].IEEE Transactions on Signal Processing,2012,60(8):4480-4485.
[6]Li C,Shen P,Liu Y.Diffusion information theoretic learning for distributed estimation over netw ork[J].IEEE Transactions on Signal Processing,2013,61(16):4011-4024.
[7]Li L,Chambers J A,et al.Distributed estimation over an adaptive incremental netw ork based on the affine projection algorithm[J].IEEE Transactions on Signal Processing,2010,58(1):151-164.
[8]李雷雷,何剑辉,张勇刚.分布式网络中的仿射投影自适应算法[J].中国传媒大学学报(自然科学版),2012,19(2):34-43.LI Lei-lei,HE Jian-hui,ZHANG Yong-gang.Affine projection algorithms over distributed w ireless netw orks[J].Journal of Communication University of China(Science and Technology),2012,19(2):34-43.(in Chinese)
[9]Sayed A H.Adaptive Filters[M].Hoboken,New Jersey:John Wiley&Sons,2008.
[10]Lee K A,Gan W S.Improving convergence of the NLMS algorithm using constrained subband updates[J].IEEE Signal Processing Letters,2004,11(9):736-739.
[11]Lee K A,Gan W S,kuo S M.Subband Adaptive Filtering Theory and Implementation[M].Hoboken,New Jersey:John Wiley&Sons,2009.
[12]倪锦根,商慧亮,李锋.归一化子带自适应滤波器步长控制[J].电子与信息学报,2009,31(11):2606-2609.NI Jin-gen,LI Feng,SHANG Hui-liang.Step-size control for the normalized subband adaptive filter[J].Journal of Electronics&Information Technology,2009,31(11):2606-2609.(in Chinese)
[13]Ni J,Li F.A variable step-size matrix normalized subband adaptive filter[J].IEEE Trans on Audio,Speech,and Language Processing,2010,18(6):1290-1299.
[14]Yang F,Wu M,et al.An improved multiband-structured subband adaptive filter algorithm[J].IEEE Signal Processing Letters,2012,19(10):647-650.
[15]Ni J,Chen X,Yang J.Two variants of the sign subband adaptive filter w ith improved convergence rate[J].Signal Processing,2014,96(B):325-331.
[16]Tu S Y,Sayed A H.Mobile adaptive networks[J].IEEE Journal of Selected Topics in Signal Processing,2011,5(4):649-664.
[17]Quan Z,Zhang W,et al.Optimal spectral feature detection for spectrum sensing at very low SNR[J].IEEE Transactions on Communications,2011,59(1):201-212.
[18]Shin Y J,Sayed A H,Shen X.Adaptive models for gene netw orks[J].PLo S ONE,2012,7(2):e31657.
[19]Lorenzo P Di,Barbarossa S,Sayed A H.Distributed spectrum estimation for small cell netw orks based on sparse diffusion adaptation[J].IEEE Signal Processing Letters,2013,20(12):1261-1265.
[20]Lorenzo P Di,Barbarossa S,Sayed A H.Bio-inspired decentralized radio access based on sw arming mechanisms overadaptive netw orks[J].IEEE Transactions on Signal Processing,2013,61(12):3183-3197.