nuralbiascompensation.m

一个采用神经网络算法进行光纤陀螺误差补偿的例子

close all
clear
clc
tic
format long
K=0.02382062019309;         %标度因数;
v0=0;                       %转台角速度
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%读数据
disp('读数据.....')

fid=fopen('F:\陀螺测试\VG951测试数据\变温情况下零偏实验数据\V1.T.dat');%
T_data=fread(fid,'double');fclose(fid);
% T_data=T_data(1:1000); %温度的平均值
fid=fopen('F:\陀螺测试\VG951测试数据\变温情况下零偏实验数据\V1.V.dat');%
v_data=fread(fid,'double');fclose(fid);
v_data=v_data(1:1000);
figure(1);
plot(T_data);
figure(2);
plot(v_data);

t=(1/300:1/300:100)';
disp('data loaded................');
figure(3);
subplot(4,2,1);
plot(t,data);
title('原始信号');
xlabel('time (s)')
%零均值检验及零均值化*******************************************3
data0=zerocheck(data);
subplot(4,2,2);
plot(t,data,t,data0,'r');
title('零均值化信号');
xlabel('time (s)')
%奇点检验及奇点删除********************************************3
[data1,s_point]=singularityCheck(data0,b);
subplot(4,2,3);                                                 
plot(t,data1,s_point(1,:)/300,s_point(2,:),'ro');
title('去奇点信号');
xlabel('time (s)')

% c_data=v_data(501:1000);
% v_data=v_data(1:500);
% v_data=v_data/max(abs(v_data));
% T=v0.*ones(size(v_data));
% 
% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
% net=newff([-1,1],[2,5,1],{'tansig','tansig','purelin'},'traingdx');
% net.trainParam.goal=1e-5;
% net.trainParam.epochs=3000;
% net=init(net);
% [net,tr]=train(net,v_data',T');
% 
% y=sim(net,v_data');
% figure;
% plot(v_data,'r');
% hold;
% plot(y);
% hold off;