function [ladderVector,...
kappaVector,...
posterioriErrorMatrix] = LRLS_pos_ErrorFeedback(desired,input,S)
% LRLS_pos_ErrorFeedback.m
% Implements the Error-Feedback LRLS algorithm based on a posteriori errors.
% (Algorithm 7.3 - book: Adaptive Filtering: Algorithms and Practical
% Implementation, Diniz)
%
% Syntax:
% [ladderVector,kappaVector,posterioriErrorMatrix] = LRLS_pos_ErrorFeedback(desired,input,S)
%
% Input Arguments:
% . desired : Desired Signal. (ROW Vector)
% . input : Signal fed into the adaptive filter. (ROW Vector)
% . S : Structure with the following fields
% - lambda : Forgetting factor. (0 << lambda < 1)
% - nSectionsLattice : Number of lattice sections, refered as N in the textbook.
% - epsilon : Initilization of xiMin_backward and xiMin_forward.
%
% Output Arguments:
% . ladderVector : Store the ladder coefficients for each iteration, refered
% as v in the textbook.
% (Coefficients at one iteration are column vector)
% . kappaVector : Store the reflection coefficients for each iteration, considering
% only the forward part.
% (Coefficients at one iteration are column vector)
% . posterioriErrorMatrix : Store the a posteriori errors at each section of the lattice for
% each iteration.
% (The errors at one iteration are column vectors)
%
% Authors:
% . Guilherme de Oliveira Pinto - guilhermepinto7@gmail.com & guilherme@lps.ufrj.br
% . Markus VinÃcius Santos Lima - mvsl20@gmailcom & markus@lps.ufrj.br
% . Wallace Alves Martins - wallace.wam@gmail.com & wallace@lps.ufrj.br
% . Luiz Wagner Pereira Biscainho - cpneqs@gmail.com & wagner@lps.ufrj.br
% . Paulo Sergio Ramirez Diniz - diniz@lps.ufrj.br
%
%################################################
% Data Initialization
%################################################
% Basic Procedure
nCoefficients = S.nSectionsLattice +1;
nIterations = length(desired);
% Pre Allocations
deltaVector = zeros(1, S.nSectionsLattice +1);
xiMin_f_Curr = zeros(1, S.nSectionsLattice +2);
xiMin_f_Prev = zeros(1, S.nSectionsLattice +2);
xiMin_b_Curr = zeros(1, S.nSectionsLattice +2);
xiMin_b_Prev = zeros(1, S.nSectionsLattice +2);
xiMin_b_PPrev = zeros(1, S.nSectionsLattice +2);
error_b_Curr = zeros(1, S.nSectionsLattice +2);
error_b_Prev = zeros(1, S.nSectionsLattice +2);
error_f = zeros(1, S.nSectionsLattice +2);
gammaVector_Curr = zeros(1, S.nSectionsLattice +2);
gammaVector_Prev = zeros(1, S.nSectionsLattice +2);
kappa_f = zeros(1, S.nSectionsLattice +1);
kappa_b = zeros(1, S.nSectionsLattice +1);
ladderVector = zeros(nCoefficients , nIterations);
kappaVector = zeros(nCoefficients , nIterations);
posterioriErrorMatrix = zeros(S.nSectionsLattice +2, nIterations);
%################################################
% Computations
%################################################
%%Initialize Parameters
%
kappa_f = zeros(1, S.nSectionsLattice +1);
kappa_b = zeros(1, S.nSectionsLattice +1);
deltaVector = zeros(1, S.nSectionsLattice +1);
%
gammaVector_Prev = ones(1,S.nSectionsLattice +2);
%
xiMin_f_Prev = repmat(S.epsilon,1, S.nSectionsLattice +2);
xiMin_b_Prev = repmat(S.epsilon,1, S.nSectionsLattice +2);
xiMin_b_PPrev = repmat(S.epsilon,1, S.nSectionsLattice +2);
%
error_b_Prev = zeros(1,S.nSectionsLattice +2);
%-------------------------------------------------------------
%||||||||||||||||||||||| - Main Loop - |||||||||||||||||||||||
%-------------------------------------------------------------
for it = 1:nIterations
% Set Values for Section 0(Zero)
gammaVector_Curr(1) = 1;
error_b_Curr(1) = input(it);
error_f(1) = input(it);
xiMin_f_Curr(1) = input(it)^2 + S.lambda*xiMin_f_Prev(1);
xiMin_b_Curr(1) = xiMin_f_Curr(1);
posterioriErrorMatrix(1,it) = desired(it);
% Propagate the Order Update Equations
for ot = 1:S.nSectionsLattice+1
%Delta Time Update
deltaVector(ot) = S.lambda*deltaVector(ot) + ...
(error_b_Prev(ot)*error_f(ot))/gammaVector_Prev(ot);
%Order Update Equations
gammaVector_Curr(ot+1) = gammaVector_Curr(ot) - ...
(error_b_Curr(ot)^2)/xiMin_b_Curr(ot);
kappa_f(ot) = (gammaVector_Prev(ot+1)/gammaVector_Prev(ot))*...
(kappa_f(ot) + ...
(error_b_Prev(ot)*error_f(ot))/...
(gammaVector_Prev(ot)*S.lambda*xiMin_b_PPrev(ot)));
kappa_b(ot) = (gammaVector_Curr(ot+1)/gammaVector_Prev(ot))*...
(kappa_b(ot) + ...
(error_b_Prev(ot)*error_f(ot))/...
(gammaVector_Prev(ot)*S.lambda*xiMin_f_Prev(ot)));
error_b_Curr(ot+1) = error_b_Prev(ot) -kappa_b(ot)*error_f(ot);
error_f(ot+1) = error_f(ot) -kappa_f(ot)*error_b_Prev(ot);
xiMin_f_Curr(ot+1) = xiMin_f_Curr(ot) -(deltaVector(ot)^2)/xiMin_b_Prev(ot);
xiMin_b_Curr(ot+1) = xiMin_b_Prev(ot) -(deltaVector(ot)^2)/xiMin_f_Curr(ot);
% Feedforward Filtering
ladderVector(ot,it+1) = (gammaVector_Curr(ot+1)/gammaVector_Curr(ot))*...
(ladderVector(ot,it) + ...
(posterioriErrorMatrix(ot,it)*error_b_Curr(ot))/...
(gammaVector_Curr(ot)*S.lambda*xiMin_b_Prev(ot)));
posterioriErrorMatrix(ot+1,it)=posterioriErrorMatrix(ot,it) - ...
ladderVector(ot,it+1)*error_b_Curr(ot);
kappaVector(ot,it) = kappa_f(ot);
end
%Upadate K-1 Info
gammaVector_Prev = gammaVector_Curr;
error_b_Prev = error_b_Curr;
xiMin_b_PPrev = xiMin_b_Prev;
xiMin_b_Prev = xiMin_b_Curr;
xiMin_f_Prev = xiMin_f_Curr;
end % END OF LOOP
%EOF
