【概率最小均方（PLMS）自适应滤波器】PLMS对高斯和非高斯噪声具有较强的鲁棒性（Matlab代码实现）

 💥💥💞💞欢迎来到本博客❤️❤️💥💥

🏆博主优势：🌞🌞🌞博客内容尽量做到思维缜密，逻辑清晰，为了方便读者。

⛳️座右铭：行百里者，半于九十。

📋📋📋本文目录如下：🎁🎁🎁

目录

💥1 概述

📚2 运行结果

🎉3 参考文献

🌈4 Matlab代码实现

---

💥1 概述

在最大后验估计框架下，本研究提出了概率最小均方（PLMS）自适应滤波器，用于从噪声数据中估计未知参数向量。PLMS将参数空间和信号空间相结合，将过程的概率分布的先验知识与信号中存在的证据相结合。利用核密度估计来估计先验分布，PLMS对高斯和非高斯噪声具有较强的鲁棒性。为了实现这一点，一些中间估计被缓冲，然后用于估计先验分布。尽管存在偏差补偿算法，但无需估计输入噪声方差。对PLMS进行了理论分析。系统识别和预测的模拟结果显示，PLMS在噪声稳态和非稳态环境中的性能可接受，而且优于偏差补偿和归一化的LMS算法。

📚2 运行结果

部分代码：

%% Initialization
N = 1;                         
M = 4;                        % Number of unknown parameters or Optimal Weight Vector Length
observations = 1000;          % Number of time observation
LMSiterations = observations; % Number of LMS Iteration
experiments = 100;            % Number of experiments for averaging over on.
Wo = ones(M,1)*sqrt(M);       % Optimal Weight Vector
stationary = 1;               % Choose 1 for stationary model, 0 for non-stationary model
N0 = .5;                      % Input noise variance
i = 1;
leg=[];
Xbar = [];
ix = 1:observations;
zinp = [];
for N1 = [.12 .38 1.3 ] %output noise variance
    display(strcat('N1= ', num2str(N1),'...'))
    for experiment = 1 : experiments
        %% Generate Input Signal and Desired Output
        inputNoise  = 'none';
        outputNoise = 'gaussian';
        N0 = N1;
        [X,d,w,zinp,zout,snr_inp] = createData(Wo,N,M,observations,N0,N1,stationary,inputNoise,outputNoise);
        snr_out = 10*log10(var(d-zout)/var(zout));
        %% PLMS Adaption
        muN =0.1;     % Step size
        bufferL = 30; % Buffer length
        [MSD(experiment,:),W,er(experiment,:)] = PLMS(X,d,LMSiterations,N,M,w,muN,bufferL);
        %% Performance Analysis of PLMS
        for j = 1 : observations
            nrmW(j) = norm(W(:,j)-Wo);
        end
        p = 1;
        for j = observations:-1:bufferL+1
            b(p) = sum(nrmW(j-bufferL:j-1))/nrmW(j);
            b(p) = b(p);
            p = p +1;
        end
        beta = mean(b);
        R = corr(reshape(X(1,:,:),M,observations)');
        I = eye(M);
        I2 = kron(I,I);
        f = I2-kron(I,muN*(I+R))-kron(muN*(I+R),I);
        q1 = muN*beta/bufferL;
        MSDT(experiment) = muN^2*N1^2*R(:)'*inv(I2-f-q1*I2)*I(:);%MSD Theory
    end
    %% Expectation over different experiments
    MeanMSDMS = mean(MSD);
    MeanMSDTheory = 10*log10(mean(MSDT)*ones(1,observations));
    
    %% Plot Results
    hold on
    plot(ix,10*log10(MeanMSDMS),clr1{i});
    leg{end+1} = strcat('PLMS, SNR_v=',num2str(floor(snr_out)));
    plot(ix,MeanMSDTheory,clr2{i})
    leg{end+1} = strcat('PLMS, SNR_v=',num2str(floor(snr_out)),'- Theory');
    i=i+1;
end
legend(leg,'Orientation','vertical','FontSize',7);
axis square
grid on
box on
xlabel('Iteration','fontsize',10)
ylabel('MSD(dB)','fontsize',10);

%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
% Estimation of Weights, Using PLMS Algorithm                    %
                              %
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
function [MSDSingle,W,er] = PLMS(X,d,LMSiterations,N,M,Wo,muN,bufferL)
W = zeros(M,1);
bufferW = zeros(M,bufferL);
h = 1;% .5;
%% Non-cooperative LMS
for n = 1 : LMSiterations
    for m = 1 : N
        ebar = [d(m,n)-X(m,:,n)*W(:,n)]*X(m,:,n)';
        WbarNodesingle(m,n) = norm(Wo(:,n)-W(:,n));
        %         WbarNodesingle(m,n) = norm(Wo-W(:,n));
        er2(m,n) = (d(m,n)-X(m,:,n)*W(:,n))^2;
        WBni = zeros(M,1); %similarity of wn and buffered wi
        sumB = 0; %normalizing factor
        for i = 1 : bufferL
            simWni = exp(-(bufferW(:,i)-W(:,n))'*(bufferW(:,i)-W(:,n))/h);
            weightedWi(:,i) = simWni*bufferW(:,i);
            sumB = sumB + simWni;
        end
        sumWi = sum(weightedWi./sumB,2);
        W(:,n+1) = W(:,n)+ muN*(ebar + sumWi - W(:,n));%gradiant ascend
        bufferW = add2buffer(bufferW,W(:,n+1));
    end
    %% Error analysis of Network
    MSDSingle(n)=norm(WbarNodesingle(:,n))^2; % Scalar of error
end
er = mean(er2,1);

🎉3 参考文献

文章中一些内容引自网络，会注明出处或引用为参考文献，难免有未尽之处，如有不妥，请随时联系删除。

[1]张以森.最小均方自适应滤波器[J].系统工程与电子技术, 1982(04):41-49.DOI:CNKI:SUN:XTYD.0.1982-04-004.

[2]王瑾,黄德修,元秀华.基于最小均方自适应滤波器的无线光通信接收性能分析[J].中国激光, 2006, 33(10):5.DOI:10.3321/j.issn:0258-7025.2006.10.017.

[3]王秋莎,张峥,王磊,等.最小均方自适应滤波器设计及性能影响因素分析[J].河北电力技术, 2023.

🌈4 Matlab代码实现