198号资源-源程序：（SCI论文＋程序）未知时延下无线传感器网络时钟同步算法研究-----已提供下载资源

电网论文源程序

于 2024-10-09 08:15:00 发布

阅读量411

点赞数 19

分类专栏：论文与完整源程序文章标签：算法

本文链接：https://blog.csdn.net/LIANG674027206/article/details/142756669

版权

论文与完整源程序专栏收录该内容

239 篇文章 231 订阅

订阅专栏

👆👆👆👆👆👆👆👆👆👆👆👆👆👆👆👆👆👆👆👆👆👆👆👆👆下载资源链接👆👆👆👆👆👆👆👆👆👆👆👆👆👆👆👆👆👆👆👆👆👆👆👆👆👆

《《《《《《《《更多资源还请持续关注本专栏》》》》》》》

论文与完整源程序_电网论文源程序的博客-CSDN博客https://blog.csdn.net/liang674027206/category_12531414.html

电网论文源程序-CSDN博客电网论文源程序擅长文章解读,论文与完整源程序,等方面的知识,电网论文源程序关注python,机器学习,计算机视觉,深度学习,神经网络,数据挖掘领域.https://blog.csdn.net/LIANG674027206?type=download

摘要：本文提出了三种时钟同步算法对于未知时延下的无线传感器网络(WSNs)来说派生的。它们包括最大似然估计器(MLE)、推广了Noh等人的估计量，并给出了一个新的低估计。复杂性估计器。它们对应的性能界限进行了推导和比较，并对其复杂性进行了分析。它是发现MLE以最低的价格获得了最好的性能具有高度的复杂性。对于广义形式的估计量Noh等人提出的，虽然复杂度较低，但其性能是相对于最大似然比降级。另一方面，新的所提出的估计器获得了与最大似然估计相同的性能，并且其复杂性与广义的Noh等人的估计量的版本。

部分代码展示：

%On clock synchronization algorithm for wireless sensor network under unknowm delay
%autor:Mei Leng and Yik-Chung Wu
%2016.12.23 by xu
%==========================================================================
close all;
%==========================================================================
%%
clear;
clc;
N = 11;%measures times
h = 1;
for SNR = 0:15:30
    %parameters setting
    O = 30;%simulation runs times
    H = 25;%中心节点发送时间间隔
    w_sigma = 0.3 * H;%中心节点发送时刻噪声的方差
    G = 30;%参考节点发送时间间隔
    v_sigma = 0.3 * G;%参考节点发送时刻噪声的方差
    d_lb = 0 + eps;%lower bound of fixed delay(open interval)
    d_ub = 10;%upper bound of fixed delay
    beta1_lb = 0.9;%lower bound of relative skew
    beta1_ub = 1.1;%upper bound of relative skew
    beta0_lb = -10;%lower bound of relative offset
    beta0_ub = 10;%upper bound of  relative offset
    sigma = (H^2+G^2)/(power(10,SNR/10));%variance of random delay Xi and Yi
    k = 1;
    for alpha = 1:10%measures times
        %time stamps
        T_1 = zeros(N,O);
        T_2 = zeros(N,O);
        T_3 = zeros(N,O);
        T_4 = zeros(N,O);
        d = zeros(1,O);%real fixed delay
        beta1 = zeros(1,O);%real clock skew
        beta0 = zeros(1,O);%real clock offset
        beta1_mle = zeros(1,O);%estimates clock skew
        beta0_mle = zeros(1,O);%estimates clock offset
        for o = 1:O
            d(1,o) = unifrnd(d_lb,d_ub);
            beta1(1,o) = unifrnd(beta1_lb,beta1_ub);
            beta0(1,o) = unifrnd(beta0_lb,beta0_ub);
            %produce measures
            for i = 1:N
                T_1(i,o) = i * H + normrnd(0,sqrt(w_sigma));
                T_3(i,o) = i * G + normrnd(0,sqrt(v_sigma));
                T_2(i,o) = beta1(1,o) * T_1(i,o) + beta0(1,o) + beta1(1,o) * (d(1,o) + normrnd(0,sqrt(sigma)));
                T_4(i,o) = (T_3(i,o) - beta0(1,o)) / beta1(1,o) + d(1,o) + normrnd(0,sqrt(sigma));
            end
            %MLLE <2>alpha= alpha star 
            %produce alpha star
            j = mod(N,3);
            n = (N - j)/3;
            alpha_star_star = 2*n+ceil(j/2);
            for alpha_star = 1:10
                for m = 1:N-alpha_star
                    D_1(m,alpha_star,o) = T_1(alpha_star+m,o)-T_1(m,o);
                    D_2(m,alpha_star,o) = T_2(alpha_star+m,o)-T_2(m,o);
                    D_3(m,alpha_star,o) = T_3(alpha_star+m,o)-T_3(m,o);
                    D_4(m,alpha_star,o) = T_4(alpha_star+m,o)-T_4(m,o);
                end
                beta1_mlle_PB2(alpha_star,o) = 2*sigma*beta1(1,o)^4/...
                    (sum(beta1(1,o)^2*D_1(:,alpha_star,o).^2+D_3(:,alpha_star,o).^2+6*beta1(1,o)^2*sigma));
            end
        end
        %MLLE  <2>alpha = alpha star
        beta1_mlle_PB2_mes(h,k) = sum(beta1_mlle_PB2(k,:),2)/O;
        k = k + 1;
    end
    figure(2)
    semilogy(1:10,beta1_mlle_PB2_mes(h,:),'ok-');
    hold on;
    semilogy(alpha_star_star,beta1_mlle_PB2_mes(h,alpha_star_star),'rx','MarkerSize',20);
    gtext(['SNR = ',num2str(SNR),' bB']);
    h=h+1;
end

效果展示：