最后,我弄清楚了如何在Matlab中使用ngreyest工具箱。这是对我有用的代码:
M = csvread('data/all/data3.txt');
u = [M(:,5), M(:,6)];
y = [M(:,4)* 1/10 * 3.1415/180, M(:,3) * 90/1000 * 3.1415/180 , M(:,2)];
base_elevationInit = y(1,1);
base_pitchInit = y(1,2);
base_travelInit = y(1,3);
%intial guess for the parameters
par = {-1.0000 -2.4000 -0.0943 0.1200 0.1200 -2.5000 -0.0200 0.2 2.1000 10.0000};
order = [3,2,6]; %[Ny Nu Nx]
initialStates =[base_elevationInit, base_pitchInit, base_travelInit, 0, 0, 0]';
Ts = 0;
m = idnlgrey('quan_model_nl',order, par, initialStates, Ts)
m.Parameters(1).Fixed = true;
m.Parameters(2).Fixed = true;
m.Parameters(8).Fixed = true;
m.Parameters(4).Fixed = true;
m.Parameters(5).Fixed = true;
m.Parameters(6).Fixed = true;
m.Parameters(9).Fixed = true;
data = iddata(y,u,0.05);
opt = nlgreyestOptions;
opt.Display = 'on';
opt.SearchOption.MaxIter = 5;
% opt.SearchMethod =
m_est = nlgreyest(data, m, opt)
params = [m_est.Parameters(1).Value m_est.Parameters(2)
我的模型函数必须保存在与上一个脚本位于同一文件夹中的名为 quan_model_nl.m 的文件中。
function [dx,y] = quan_model_nl(t, x, u, p1, p2, p3, p4, p5, p6, p7, p8, p9, p10, varargin)
F = [ x(4);
x(5);
x(6);
p1*cos(x(1))+ p2*sin(x(1)) + p3*x(6);
p5*sin(x(2)) + p4*cos(x(2))+ p6*x(5);
p7*x(6);
];
G = [
0 0 ;
0 0 ;
0 0 ;
p8*cos(x(2)) p8*cos(x(2)) ;
p9 -p9 ;
p10*sin(x(2)) p10*sin(x(2)) ;
];
C = [
1,0,0,0,0,0;
0,1,0,0,0,0;
0,0,1,0,0,0;
];
dx = F + G * u';
y = C * x ;
end