FUNCTION hybrid() implement the hybrid filter combining the CLMS
algorithm and ACLMS algorithm.
Based on the paper "Collaborative adaptive filtering in the complex domain"
IEEE Workshop on Machine Learning for Signal Processing, 421-425, 2008.
INPUT:
x: input signal
N: filter length of CLMS (that of ACLMS is N1 * 2)
mu1: step-size of CLMS
mu2: step-size of ACLMS
lambda: mixing parameter
mu3: step-size of adaptation of lambda
OUTPUT:
y: filter output
Complex Valued Nonlinear Adaptive Filtering toolbox for MATLAB
Supplementary to the book:
"Complex Valued Nonlinear Adaptive Filters: Noncircularity, Widely Linear and Neural Models"
by Danilo P. Mandic and Vanessa Su Lee Goh
(c) Copyright Danilo P. Mandic 2009
http://www.commsp.ee.ic.ac.uk/~mandic
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
This program is free software; you can redistribute it and/or modify
it under the terms of the GNU General Public License as published by
the Free Software Foundation; either version 2 of the License, or
(at your option) any later version.
This program is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
GNU General Public License for more details.
You can obtain a copy of the GNU General Public License from
http://www.gnu.org/copyleft/gpl.html or by writing to
Free Software Foundation, Inc.,675 Mass Ave, Cambridge, MA 02139, USA.
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
...........................................0001 % FUNCTION hybrid() implement the hybrid filter combining the CLMS 0002 % algorithm and ACLMS algorithm. 0003 % 0004 % Based on the paper "Collaborative adaptive filtering in the complex domain" 0005 % IEEE Workshop on Machine Learning for Signal Processing, 421-425, 2008. 0006 % 0007 % INPUT: 0008 % x: input signal 0009 % N: filter length of CLMS (that of ACLMS is N1 * 2) 0010 % mu1: step-size of CLMS 0011 % mu2: step-size of ACLMS 0012 % lambda: mixing parameter 0013 % mu3: step-size of adaptation of lambda 0014 % 0015 % OUTPUT: 0016 % y: filter output 0017 % 0018 % 0019 % Complex Valued Nonlinear Adaptive Filtering toolbox for MATLAB 0020 % Supplementary to the book: 0021 % 0022 % "Complex Valued Nonlinear Adaptive Filters: Noncircularity, Widely Linear and Neural Models" 0023 % by Danilo P. Mandic and Vanessa Su Lee Goh 0024 % 0025 % (c) Copyright Danilo P. Mandic 2009 0026 % http://www.commsp.ee.ic.ac.uk/~mandic 0027 % 0028 % %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% 0029 % This program is free software; you can redistribute it and/or modify 0030 % it under the terms of the GNU General Public License as published by 0031 % the Free Software Foundation; either version 2 of the License, or 0032 % (at your option) any later version. 0033 % 0034 % This program is distributed in the hope that it will be useful, 0035 % but WITHOUT ANY WARRANTY; without even the implied warranty of 0036 % MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the 0037 % GNU General Public License for more details. 0038 % 0039 % You can obtain a copy of the GNU General Public License from 0040 % http://www.gnu.org/copyleft/gpl.html or by writing to 0041 % Free Software Foundation, Inc.,675 Mass Ave, Cambridge, MA 02139, USA. 0042 % %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% 0043 % ........................................... 0044 function y = hybrid(x,N,mu1,mu2,lambda,mu3) 0045 0046 M = 1;% prediction horizon 0047 L = length(x)-M; % 0048 filterinput = zeros(N,L); 0049 WCLMS = zeros(N,1); % weight vector of subfilter 1 0050 eCLMS = zeros(1,L); % error of subfilter 1 0051 suboutput1 = zeros(1,L); % output of subfilter 1 0052 0053 WACLMS = zeros(2*N,1); % weight vector of subfilter 2 0054 eACLMS = zeros(1,L); % error of subfilter 2 0055 suboutput2 = zeros(1,L); % output of subfilter 2 0056 0057 ehybrid = zeros(1,L); 0058 filteroutput = zeros(1,L); 0059 0060 for i = 1:L 0061 for m = 1:N 0062 if (i-m+1)>0 0063 filterinput(m,i) = x(1,i-m+1); 0064 else 0065 filterinput(m,i) = 0; 0066 end 0067 end 0068 suboutput1(i) = transpose(filterinput(:,i)) * WCLMS; 0069 suboutput2(i) = transpose([filterinput(:,i);conj(filterinput(:,i))]) * WACLMS; 0070 eCLMS(i) = x(1,i+M) - suboutput1(i); 0071 eACLMS(i) = x(1,i+M) - suboutput2(i); 0072 filteroutput(i) = lambda * suboutput1(i) + (1-lambda) * suboutput2(i); 0073 ehybrid(i) = x(1,i+M) - filteroutput(i);% error(k) 0074 WCLMS = WCLMS + mu1 * eCLMS(i) * conj(filterinput(:,i));% weight update of subfilter 1 0075 WACLMS = WACLMS + mu2 * eACLMS(i) * conj([filterinput(:,i);conj(filterinput(:,i))]);% weight update of subfilter 2 0076 lambda = lambda + mu3 * real(ehybrid(i) * conj(suboutput1(i)-suboutput2(i)));% update of lambda 0077 end 0078 y = filteroutput; 0079