gusucode.com > 扩展卡尔曼滤波，粒子滤波，去偏卡尔曼滤波和循环增益尔曼滤波的源程序 > 去偏卡尔曼滤波/RMSE_err_count.m

    function [RMSE_final_view_err_x,RMSE_final_filter_err_x,RMSE_final_view_err_y,RMSE_final_filter_err_y]=RMSE_err_count(ME_temp_view_err_x,ME_temp_filter_err_x,ME_temp_view_err_y,ME_temp_filter_err_y,ME_final_view_err_x,ME_final_filter_err_x,ME_final_view_err_y,ME_final_filter_err_y)
[ROW,COL]=size(ME_temp_view_err_x);
for i=1:ROW
    RVX(i,:)=ME_temp_view_err_x(i,:)-ME_final_view_err_x;                    %RVX代表x方向单个观测误差样本减去均值的矩阵
    RFX(i,:)=ME_temp_filter_err_x(i,:)-ME_final_filter_err_x;                %RFX代表x方向单个滤波误差样本减去均值的矩阵
    RVY(i,:)=ME_temp_view_err_y(i,:)-ME_final_view_err_y;                    %RVX代表y方向单个观测误差样本减去均值的矩阵
    RFY(i,:)=ME_temp_filter_err_y(i,:)-ME_final_filter_err_y;                %RFX代表y方向单个滤波误差样本减去均值的矩阵
end
for i=1:COL
    RMSE_final_view_err_x(i)=sqrt(sum(RVX(:,i).*RVX(:,i))/(ROW-1));                   %x方向的观测均方误差
    RMSE_final_filter_err_x(i)=sqrt(sum(RFX(:,i).*RFX(:,i))/(ROW-1));                  %x方向的滤波均方误差
    RMSE_final_view_err_y(i)=sqrt(sum(RVY(:,i).*RVY(:,i))/(ROW-1));                   %y方向的观测均方误差
    RMSE_final_filter_err_y(i)=sqrt(sum(RFY(:,i).*RFY(:,i))/(ROW-1));                  %y方向的滤波均方误差
end