-
Notifications
You must be signed in to change notification settings - Fork 0
/
main_fit_data_neural_net.m
324 lines (286 loc) · 10.2 KB
/
main_fit_data_neural_net.m
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
%%%%% new: TE dividing line, one fit for each control %%%%%%%
%function [control_fit_theta,control_fit_Pc,control_fit_Pt,net_theta,net_Pc,net_Pt] = main_fit_data_neural_net(data_file,net_theta,net_Pc,net_Pt,XX,YY)
function [net_theta,net_Pc,net_Pt,cluster_data,robot] = main_fit_data_neural_net(data_file)
if (nargin==0)
clc
close all
clear all
end
Fontsize = 12;
cluster_data = 1; %0 no clustering, 1 for clustering
fig_no = 1;
if (nargin==0)
%clear variables
% 1 file
%data_file = {'data_lyap1_09_50_14.mat'};
%data_file = {'data_lyap1_09_50_14.mat';...
% 'data_lyap1_09_30_15.mat'};
% data_file = {'data_lyap1_09_50_14.mat';...
% 'data_lyap1_09_30_15.mat'; ...
% 'data_lyap1_09_20_13.mat';...
% 'data_lyap1_09_40_14.mat'};
data_file = {'data_lyap1_09_20_13.mat';...
'data_lyap1_09_20_16.mat'; ...
'data_lyap1_09_30_15.mat';...
'data_lyap1_09_40_14.mat';...
'data_lyap1_09_40_16.mat';...
'data_lyap1_09_50_14.mat';...
'data_lyap1_09_50_16.mat'};
% data_file = {'data/data_lyap1_09_30_14_90.mat'};
else
%load(data_file);
end
for iiii=1:3
n1 = 1;
data_type = iiii;
rand('seed',0);
%%%%%%%%%%%%%% 1. Pull out the data from the files %%%%%%%%%%%%%%
for pppp=1:length(data_file)
%pppp=1;
load(data_file{pppp});
%%%%% needs to be rewritten (PAB47)
% if (nargin>1) %overwrite X and Y (do this later)
% X = XX;
% Y = YY;
% end
if (data_type==1)
string_title = 'Control theta';
data = control_theta;
%all_data(pppp).control_theta = control_theta;
elseif (data_type ==2)
string_title = 'Control Pc';
data = control_Pc;
%all_data(pppp).control_Pc = control_Pc;
elseif (data_type ==3)
string_title = 'Control Pt';
data = control_Pt;
%all_data(pppp).control_Pt = control_Pt;
else
error('data_type can be 1, 2, or 3');
end
all_data(pppp).data = data;
all_data(pppp).X = X;
all_data(pppp).Y = Y;
all_data(pppp).zstar = robot.zstar;
all_data(pppp).theta0 = theta0;
TE0 = 0.5*robot.m*robot.zstar(1)*robot.zstar(1) + robot.m*robot.g*robot.zstar(2);
all_data(pppp).TE0 = TE0;
[mmm,nnn] = size(X);
ppp = mmm*nnn;
%if (data_type == 2 || data_type==3)
% if (nargin<2)
% for i=1:mmm
% for j=1:nnn
% if (data(i,j)>8000)
% data(i,j) = NaN;
% %graphingZ1(i,j) = NaN;
% end
% end
% end
% end
k1 = 1;
if (cluster_data == 1 && data_type ~= 1)
for i=1:mmm
for j=1:nnn
vx_temp = X(i,j);
h_temp = Y(i,j);
TE_temp = 0.5*robot.m*vx_temp*vx_temp + robot.m*robot.g*h_temp;
%del_vx = vx_temp - robot.zstar(1);
%del_h = h_temp - robot.zstar(2);
%if (TE>TE_temp)
%if (TE_temp<controls(7))
if (data_type == 2) %Pc
if (TE_temp>TE0) %only save TE > TE0 for fitting
netrainX(n1,1) = X(i,j); % F(i,j) = 0;
netrainX(n1,2) = Y(i,j);
netrainX(n1,3) = robot.zstar(1); %velocity
netrainX(n1,4) = robot.zstar(2); %apex height
all_data(pppp).netrainX(k1,1) = X(i,j);
all_data(pppp).netrainX(k1,2) = Y(i,j);
all_data(pppp).netrainX(k1,3) = robot.zstar(1);
all_data(pppp).netrainX(k1,4) = robot.zstar(2);
%if (nargin<2)
if (exitflag(i,j) == 1) % || exitflag(i,j) == 41)
netrainY1(n1,1) = data(i,j);
else
netrainY1(n1,1) = NaN;
end
%end
all_data(pppp).index(k1,:) = [i,j];
k1 = k1+1;
index(n1,:) = [i,j];
n1 = n1+1;
end
end
if (data_type == 3) %Pt
if (TE_temp<TE0) %only save TE < TE0 for fitting
netrainX(n1,1) = X(i,j);
netrainX(n1,2) = Y(i,j);
netrainX(n1,3) = robot.zstar(1); %velocity
netrainX(n1,4) = robot.zstar(2); %apex height
all_data(pppp).netrainX(k1,1) = X(i,j);
all_data(pppp).netrainX(k1,2) = Y(i,j);
all_data(pppp).netrainX(k1,3) = robot.zstar(1);
all_data(pppp).netrainX(k1,4) = robot.zstar(2);
%if (nargin<2)
if (exitflag(i,j) == 1) % || exitflag(i,j) == 41)
netrainY1(n1,1) = data(i,j);
else
netrainY1(n1,1) = NaN;
end
%end
all_data(pppp).index(k1,:) = [i,j];
k1 = k1+1;
index(n1,:) = [i j];
n1 = n1+1;
end
end
end
end
else
for i=1:mmm
for j=1:nnn
netrainX(n1,1) = X(i,j);
netrainX(n1,2) = Y(i,j);
netrainX(n1,3) = robot.zstar(1); %velocity
netrainX(n1,4) = robot.zstar(2); %apex height
all_data(pppp).netrainX(k1,1) = X(i,j);
all_data(pppp).netrainX(k1,2) = Y(i,j);
all_data(pppp).netrainX(k1,3) = robot.zstar(1);
all_data(pppp).netrainX(k1,4) = robot.zstar(2);
%if (nargin<2)
if (exitflag(i,j) == 1) % || exitflag(i,j) == 41)
netrainY1(n1,1) = data(i,j);
else
netrainY1(n1,1) = NaN;
end
%end
k1 = k1+1;
n1 = n1 + 1;
end
end
end
end
%%%%%%%%% 2. Data fitting %%%%%%%%%%%%%%%%%%%%%%
%net1 = fitnet([10,8,6]);
%if (nargin<2)
net1 = fitnet(12);
%view(net1)
net1 = train(net1,netrainX',netrainY1');
%else
% if (data_type == 1)
% net1 = net_theta;
% elseif (data_type == 2)
% net1 = net_Pc;
% else
% net1 = net_Pt;
% end
%end
%%%%%%%%%%% 3 a. Forecasting (errors) %%%%%%%%%%%%%%
%mean_sq_error = mse(net1,netrainX',netrainY1')
%mean_sq_error/ppp
% forecast(1).Y = net1(netestX')';
sum_ppp = 0;
sum = 0;
for pppp=1:length(data_file)
%forecast(1).Y = net1(netrainX')';
forecast(pppp).Y = net1(all_data(pppp).netrainX')';
%size(all_data(pppp).netrainX)
%if (nargin<2)
mean = 0;
if (data_type == 1) %for control_theta, mean is not zero
mean = all_data(pppp).theta0;
end
for i=1:length(forecast(pppp).Y)
sum = sum+(forecast(pppp).Y(i) - mean)^2;
end
sum_ppp = sum_ppp + ppp;
%end
end
disp(string_title);
disp('mse is');
disp(sum/sum_ppp)
%%%%%%% 3 b. Graphing the forecasts (results) %%%%%%%%
indices = 1:1:length(data_file);
if(data_type==2)
indices = length(data_file):-1:1;
end
%for pppp=1:length(data_file)
for pppp=indices
nx=1;
graphingX = all_data(pppp).X;
graphingY = all_data(pppp).Y;
forecastedData1 = zeros(mmm,nnn);
%if (nargin<2)
graphingZ1 = all_data(pppp).data;
%end
if (cluster_data == 1 && data_type ~= 1)
for px = 1:length(all_data(pppp).index)
iii = all_data(pppp).index(px,1); jjj = all_data(pppp).index(px,2);
if (forecast(pppp).Y(px,1)<0)
forecastedData1(iii,jjj) = 0;
else
forecastedData1(iii,jjj) = forecast(pppp).Y(px,1);
end
%disp([graphingX(iii,jjj) graphingX(iii,jjj) graphingZ1(iii,jjj)/1000 data(iii,jjj)/1000])
end
else
for px=1:mmm
for qx=1:nnn
%graphingX(px,qx) = netrainX(nx,1);
%graphingY(px,qx) = netrainX(nx,2);
forecastedData1(px,qx) = forecast(pppp).Y(nx,1);
%graphingZ1(px,qx) = netrainY1(nx,1);
nx = nx + 1;
end
end
end
if (nargin==0)
if (data_type==1)
figure(1);
elseif (data_type==2)
figure(2);
else
figure(3);
end
%fig_no = fig_no +1;
subplot(2,1,2);
contourf(graphingX,graphingY,forecastedData1,'ShowText','on'); hold on;
%colormap(flipud(autumn));
title([string_title,' predict'],'Fontsize',Fontsize);
subplot(2,1,1);
contourf(graphingX,graphingY,graphingZ1,'ShowText','on'); hold on;
%colormap(flipud(autumn));
title([string_title,' actual'],'Fontsize',Fontsize);
xlabel('vx','Fontsize',Fontsize); ylabel('y','Fontsize',Fontsize);
for iijj=1:2
subplot(2,1,iijj)
if(data_type==1)
colormap('cool');
elseif (data_type==2)
colormap(flipud(autumn));
else
colormap(flipud(summer));
end
end
end
%%%%% needs to be rewritten (PAB47)
% if (data_type==1)
% control_fit_theta = forecastedData1;
% elseif (data_type == 2)
% control_fit_Pc = forecastedData1;
% else
% control_fit_Pt = forecastedData1;
% end
end
if (data_type == 1)
net_theta = net1;
elseif (data_type == 2)
net_Pc = net1;
else
net_Pt = net1;
end
clear net1
clear netrainX netrainY1 forecast forecastedData1 index all_data
rmfield(robot,{'zstar','control'});
end