-
Notifications
You must be signed in to change notification settings - Fork 1
/
main_lorenz.py
330 lines (260 loc) · 15.1 KB
/
main_lorenz.py
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
325
import torch
torch.pi = torch.acos(torch.zeros(1)).item() * 2 # which is 3.1415927410125732
import torch.nn as nn
from EKF_test import EKFTest
from UKF_test import UKFTest
from PF_test import PFTest
from Extended_sysmdl import SystemModel
from Extended_data import DataGen,DataLoader,DataLoader_GPU, Decimate_and_perturbate_Data,Short_Traj_Split
from Extended_data import N_E, N_CV, N_T
from Pipeline_EKF import Pipeline_EKF
from Extended_KalmanNet_nn import KalmanNetNN
from datetime import datetime
from Plot import Plot_extended as Plot
from filing_paths import path_model
import sys
sys.path.insert(1, path_model)
from parameters import T, T_test, m1x_0, m2x_0, m, n,delta_t_gen,delta_t
from model import f, h, fInacc, hInacc, fRotate, h_nonlinear
if torch.cuda.is_available():
dev = torch.device("cuda:0") # you can continue going on here, like cuda:1 cuda:2....etc.
torch.set_default_tensor_type('torch.cuda.FloatTensor')
print("Running on the GPU")
else:
dev = torch.device("cpu")
print("Running on the CPU")
print("Pipeline Start")
################
### Get Time ###
################
today = datetime.today()
now = datetime.now()
strToday = today.strftime("%m.%d.%y")
strNow = now.strftime("%H:%M:%S")
strTime = strToday + "_" + strNow
print("Current Time =", strTime)
#########################
### Set parameters ###
#########################
offset = 0
chop = False
DatafolderName = 'Simulations/Lorenz_Atractor/data/T2000_NT100' + '/'
# data_gen = 'data_gen.pt'
# data_gen_file = torch.load(DatafolderName+data_gen, map_location=dev)
# [true_sequence] = data_gen_file['All Data']
r2 = torch.tensor([1,0.1,0.01,1e-3,1e-4])
# r2 = torch.tensor([100, 10, 1, 0.1, 0.01])
r = torch.sqrt(r2)
vdB = -20 # ratio v=q2/r2
v = 10**(vdB/10)
q2 = torch.mul(v,r2)
q = torch.sqrt(q2)
### q and r searched for filters
r2searchdB = torch.tensor([-5,0,5])
rsearch = torch.sqrt(10**(-r2searchdB/10))
q2searchdB = torch.tensor([20,15,10])
qsearch = torch.sqrt(10**(-q2searchdB/10))
### q and r optimized for filters
r2optdB = torch.tensor([3.0103])
ropt = torch.sqrt(10**(-r2optdB/10))
r2optdB_partial = torch.tensor([3.0103])
ropt_partial = torch.sqrt(10**(-r2optdB_partial/10))
q2optdB = torch.tensor([18.2391,28.2391,38.2391,48,55])
qopt = torch.sqrt(10**(-q2optdB/10))
q2optdB_partial = torch.tensor([18.2391,28.2391,38.2391,48,55])
qopt_partial = torch.sqrt(10**(-q2optdB_partial/10))
# traj_resultName = ['traj_lor_KNetFull_rq1030_T2000_NT100.pt']#,'partial_lor_r4.pt','partial_lor_r5.pt','partial_lor_r6.pt']
dataFileName = ['data_lor_v20_rq020_T2000.pt','data_lor_v20_rq1030_T2000.pt','data_lor_v20_rq2040_T2000.pt','data_lor_v20_rq3050_T2000.pt','data_lor_v20_rq4060_T2000.pt']# for T=2000
# dataFileName = ['data_lor_v20_rq020_T1000_NT100.pt','data_lor_v20_rq1030_T1000_NT100.pt','data_lor_v20_rq2040_T1000_NT100.pt','data_lor_v20_rq3050_T1000_NT100.pt']# for T=1000
EKFResultName = ['EKF_rq020_T2000','EKF_rq1030_T2000','EKF_rq2040_T2000','EKF_rq3050_T2000','EKF_rq4060_T2000']
UKFResultName = ['UKF_rq020_T2000','UKF_rq1030_T2000','UKF_rq2040_T2000','UKF_rq3050_T2000','UKF_rq4060_T2000']
PFResultName = ['PF_rq020_T2000','PF_rq1030_T2000','PF_rq2040_T2000','PF_rq3050_T2000','PF_rq4060_T2000']
for index in range(0, len(r)):
print("1/r2 [dB]: ", 10 * torch.log10(1/r[index]**2))
print("1/q2 [dB]: ", 10 * torch.log10(1/q[index]**2))
#############################
### Prepare System Models ###
#############################
sys_model = SystemModel(f, q[index], h, r[index], T, T_test, m, n,"Lor")
sys_model.InitSequence(m1x_0, m2x_0)
sys_model_optq = SystemModel(f, qopt[index], h, r[index], T, T_test, m, n,"Lor")
sys_model_optq.InitSequence(m1x_0, m2x_0)
sys_model_partialf_optq = SystemModel(fInacc, qopt_partial[index], h, r[index], T, T_test, m, n,"Lor")
sys_model_partialf_optq.InitSequence(m1x_0, m2x_0)
# sys_model_partialh = SystemModel(f, q[index], h_nonlinear, r[index], T, T_test, m, n,"Lor")
# sys_model_partialh.InitSequence(m1x_0, m2x_0)
# sys_model_partialh_optr = SystemModel(f, q[index], h_nonlinear, ropt, T, T_test, m, n,'lor')
# sys_model_partialh_optr.InitSequence(m1x_0, m2x_0)
#################################
### Generate and load DT data ###
#################################
# print("Start Data Gen")
# DataGen(sys_model, DatafolderName + dataFileName[index], T, T_test,randomInit=False)
print("Data Load")
print(dataFileName[index])
[train_input_long,train_target_long, cv_input, cv_target, test_input, test_target] = torch.load(DatafolderName + dataFileName[index],map_location=dev)
if chop:
print("chop training data")
[train_target, train_input] = Short_Traj_Split(train_target_long, train_input_long, T)
# [cv_target, cv_input] = Short_Traj_Split(cv_target, cv_input, T)
else:
print("no chopping")
train_target = train_target_long[:,:,0:T]
train_input = train_input_long[:,:,0:T]
# cv_target = cv_target[:,:,0:T]
# cv_input = cv_input[:,:,0:T]
print("trainset size:",train_target.size())
print("cvset size:",cv_target.size())
print("testset size:",test_target.size())
"""
############################################################
### Generate and load data for decimation case (chopped) ###
############################################################
print("Data Gen")
[test_target, test_input] = Decimate_and_perturbate_Data(true_sequence, delta_t_gen, delta_t, N_T, h, r[rindex], offset)
print(test_target.size())
[train_target_long, train_input_long] = Decimate_and_perturbate_Data(true_sequence, delta_t_gen, delta_t, N_E, h, r[rindex], offset)
[cv_target_long, cv_input_long] = Decimate_and_perturbate_Data(true_sequence, delta_t_gen, delta_t, N_CV, h, r[rindex], offset)
[train_target, train_input] = Short_Traj_Split(train_target_long, train_input_long, T)
[cv_target, cv_input] = Short_Traj_Split(cv_target_long, cv_input_long, T)
"""
################################
### Evaluate EKF, UKF and PF ###
################################
### grid search of opt q for benchmarks
# for searchindex in range(0, len(qsearch)):
# print("\n Searched optimal 1/q2 [dB]: ", 10 * torch.log10(1/qsearch[searchindex]**2))
# sys_model_searchq = SystemModel(f, qsearch[searchindex], h, r[index], T, T_test, m, n,"Lor")
# sys_model_searchq.InitSequence(m1x_0, m2x_0)
# print("Evaluate EKF true")
# [MSE_EKF_linear_arr, MSE_EKF_linear_avg, MSE_EKF_dB_avg, EKF_KG_array, EKF_out] = EKFTest(sys_model_searchq, test_input, test_target)
# print("Evaluate UKF True")
# [MSE_UKF_linear_arr, MSE_UKF_linear_avg, MSE_UKF_dB_avg, UKF_out] = UKFTest(sys_model_searchq, test_input, test_target)
# print("Evaluate PF True")
# [MSE_PF_linear_arr, MSE_PF_linear_avg, MSE_PF_dB_avg, PF_out] = PFTest(sys_model_searchq, test_input, test_target)
# # Filters only have partial info of process model
# sys_model_partialf_searchq = SystemModel(fInacc, qsearch[searchindex], h, r[index], T, T_test, m, n,'lor')
# sys_model_partialf_searchq.InitSequence(m1x_0, m2x_0)
# print("Evaluate EKF Partial")
# [MSE_EKF_linear_arr_partial, MSE_EKF_linear_avg_partial, MSE_EKF_dB_avg_partial, EKF_KG_array_partial, EKF_out_partial] = EKFTest(sys_model_partialf_searchq, test_input, test_target)
# print("Evaluate UKF Partial")
# [MSE_UKF_linear_arr_partial, MSE_UKF_linear_avg_partial, MSE_UKF_dB_avg_partial, UKF_out_partial] = UKFTest(sys_model_partialf_searchq, test_input, test_target)
# print("Evaluate PF Partial")
# [MSE_PF_linear_arr_partial, MSE_PF_linear_avg_partial, MSE_PF_dB_avg_partial, PF_out_partial] = PFTest(sys_model_partialf_searchq, test_input, test_target)
### grid search of opt r for benchmarks
# for searchindex in range(0, len(rsearch)):
# print("\n Searched optimal 1/r2 [dB]: ", 10 * torch.log10(1/rsearch[searchindex]**2))
# sys_model_searchr = SystemModel(f, q[index], h, rsearch[searchindex], T, T_test, m, n,"Lor")
# sys_model_searchr.InitSequence(m1x_0, m2x_0)
# print("Evaluate EKF true")
# [MSE_EKF_linear_arr, MSE_EKF_linear_avg, MSE_EKF_dB_avg, EKF_KG_array, EKF_out] = EKFTest(sys_model_searchr, test_input, test_target)
# print("Evaluate UKF True")
# [MSE_UKF_linear_arr, MSE_UKF_linear_avg, MSE_UKF_dB_avg, UKF_out] = UKFTest(sys_model_searchr, test_input, test_target)
# print("Evaluate PF True")
# [MSE_PF_linear_arr, MSE_PF_linear_avg, MSE_PF_dB_avg, PF_out] = PFTest(sys_model_searchr, test_input, test_target)
# # Filters only have partial info of observation model
# sys_model_partialh_searchr = SystemModel(f, q[index], hInacc, rsearch[searchindex], T, T_test, m, n,"Lor")
# sys_model_partialh_searchr.InitSequence(m1x_0, m2x_0)
# print("Evaluate EKF Partial")
# [MSE_EKF_linear_arr_partial, MSE_EKF_linear_avg_partial, MSE_EKF_dB_avg_partial, EKF_KG_array_partial, EKF_out_partial] = EKFTest( sys_model_partialh_searchr, test_input, test_target)
# print("Evaluate UKF Partial")
# [MSE_UKF_linear_arr_partial, MSE_UKF_linear_avg_partial, MSE_UKF_dB_avg_partial, UKF_out_partial] = UKFTest(sys_model_partialh_searchr, test_input, test_target)
# print("Evaluate PF Partial")
# [MSE_PF_linear_arr_partial, MSE_PF_linear_avg_partial, MSE_PF_dB_avg_partial, PF_out_partial] = PFTest( sys_model_partialh_searchr, test_input, test_target)
print("Evaluate EKF true")
[MSE_EKF_linear_arr, MSE_EKF_linear_avg, MSE_EKF_dB_avg, EKF_KG_array, EKF_out] = EKFTest(sys_model_optq, test_input, test_target)
print("Evaluate UKF True")
[MSE_UKF_linear_arr, MSE_UKF_linear_avg, MSE_UKF_dB_avg, UKF_out] = UKFTest(sys_model_optq, test_input, test_target)
print("Evaluate PF True")
[MSE_PF_linear_arr, MSE_PF_linear_avg, MSE_PF_dB_avg, PF_out] = PFTest(sys_model_optq, test_input, test_target)
#Evaluate partial_f
print("Evaluate EKF Partial")
[MSE_EKF_linear_arr_partialf, MSE_EKF_linear_avg_partialf, MSE_EKF_dB_avg_partialf, EKF_KG_array_partialf, EKF_out_partialf] = EKFTest(sys_model_partialf_optq, test_input, test_target)
print("Evaluate UKF Partial")
[MSE_UKF_linear_arr_partialf, MSE_UKF_linear_avg_partialf, MSE_UKF_dB_avg_partialf, UKF_out_partialf] = UKFTest(sys_model_partialf_optq, test_input, test_target)
print("Evaluate PF Partial")
[MSE_PF_linear_arr_partialf, MSE_PF_linear_avg_partialf, MSE_PF_dB_avg_partialf, PF_out_partialf] = PFTest(sys_model_partialf_optq, test_input, test_target)
# Save results
FilterfolderName = 'Filters/DT case/histogram/procmis/T2000' + '/'
torch.save({'MSE_EKF_linear_arr': MSE_EKF_linear_arr,
'MSE_EKF_dB_avg': MSE_EKF_dB_avg,
'EKF_out':EKF_out,
'MSE_EKF_linear_arr_partial': MSE_EKF_linear_arr_partialf,
'MSE_EKF_dB_avg_partial': MSE_EKF_dB_avg_partialf,
'EKF_out_partial': EKF_out_partialf,
# 'MSE_EKF_linear_arr_partialh': MSE_EKF_linear_arr_partialh,
# 'MSE_EKF_dB_avg_partialh': MSE_EKF_dB_avg_partialh,
}, FilterfolderName+EKFResultName[index])
torch.save({'MSE_UKF_linear_arr': MSE_UKF_linear_arr,
'MSE_UKF_dB_avg': MSE_UKF_dB_avg,
'UKF_out':UKF_out,
'MSE_UKF_linear_arr_partialf': MSE_UKF_linear_arr_partialf,
'MSE_UKF_dB_avg_partialf': MSE_UKF_dB_avg_partialf,
'UKF_out_partialf': UKF_out_partialf,
# 'MSE_UKF_linear_arr_partialh': MSE_UKF_linear_arr_partialh,
# 'MSE_UKF_dB_avg_partialh': MSE_UKF_dB_avg_partialh,
}, FilterfolderName+UKFResultName[index])
torch.save({'MSE_PF_linear_arr': MSE_PF_linear_arr,
'MSE_PF_dB_avg': MSE_PF_dB_avg,
'PF_out':PF_out,
'MSE_PF_linear_arr_partialf': MSE_PF_linear_arr_partialf,
'MSE_PF_dB_avg_partialf': MSE_PF_dB_avg_partialf,
'PF_out_partialf': PF_out_partialf,
# 'MSE_EKF_linear_arr_partialoptr': MSE_EKF_linear_arr_partialoptr,
# 'MSE_EKF_dB_avg_partialoptr': MSE_EKF_dB_avg_partialoptr,
}, FilterfolderName+PFResultName[index])
#####################
### Evaluate KNet ###
#####################
### KNet without model mismatch
# sys_model = SystemModel(f, q[0], h, r[0], T, T_test, m, n,"Lor")# arbitary q and r
# sys_model.InitSequence(m1x_0, m2x_0)
# print("KNet with full model info")
# modelFolder = 'KNet' + '/'
# KNet_Pipeline = Pipeline_EKF(strTime, "KNet", "KalmanNet")
# KNet_Pipeline.setssModel(sys_model)
# KNet_model = KalmanNetNN()
# KNet_model.Build(sys_model)
# KNet_Pipeline.setModel(KNet_model)
# KNet_Pipeline.setTrainingParams(n_Epochs=200, n_Batch=10, learningRate=1e-3, weightDecay=1e-4)
# KNet_Pipeline.model = torch.load(modelFolder+"model_KNet.pt")
# KNet_Pipeline.NNTrain(N_E, train_input, train_target, N_CV, cv_input, cv_target)
# [KNet_MSE_test_linear_arr, KNet_MSE_test_linear_avg, KNet_MSE_test_dB_avg, KNet_test] = KNet_Pipeline.NNTest(N_T, test_input, test_target)
# KNet_Pipeline.save()
### KNet with model mismatch
# print("KNet with model mismatch")
# sys_model_partialh = SystemModel(f, q[0], hInacc, r[0], T, T_test, m, n,"Lor")# arbitary q and r
# sys_model_partialh.InitSequence(m1x_0, m2x_0)
# modelFolder = 'KNet' + '/'
# KNet_Pipeline = Pipeline_EKF(strTime, "KNet", "KNet")
# KNet_Pipeline.setssModel(sys_model_partialh)
# KNet_model = KalmanNetNN()
# KNet_model.Build(sys_model_partialh)
# KNet_Pipeline.setModel(KNet_model)
# KNet_Pipeline.setTrainingParams(n_Epochs=200, n_Batch=10, learningRate=1e-3, weightDecay=1e-4)
# # KNet_Pipeline.model = torch.load(modelFolder+"model_KNet_obsmis_rq1030_T2000.pt",map_location=dev)
# KNet_Pipeline.NNTrain(N_E, train_input, train_target, N_CV, cv_input, cv_target)
# [KNet_MSE_test_linear_arr, KNet_MSE_test_linear_avg, KNet_MSE_test_dB_avg, KNet_test] = KNet_Pipeline.NNTest(N_T, test_input, test_target)
# KNet_Pipeline.save()
# # Save trajectories
# # trajfolderName = 'KNet' + '/'
# # DataResultName = traj_resultName[rindex]
# # # EKF_sample = torch.reshape(EKF_out[0,:,:],[1,m,T_test])
# # # EKF_Partial_sample = torch.reshape(EKF_out_partial[0,:,:],[1,m,T_test])
# # # target_sample = torch.reshape(test_target[0,:,:],[1,m,T_test])
# # # input_sample = torch.reshape(test_input[0,:,:],[1,n,T_test])
# # # KNet_sample = torch.reshape(KNet_test[0,:,:],[1,m,T_test])
# # torch.save({
# # 'KNet': KNet_test,
# # }, trajfolderName+DataResultName)
# ## Save histogram
# EKFfolderName = 'KNet' + '/'
# torch.save({'MSE_EKF_linear_arr': MSE_EKF_linear_arr,
# 'MSE_EKF_dB_avg': MSE_EKF_dB_avg,
# 'MSE_EKF_linear_arr_partial': MSE_EKF_linear_arr_partial,
# 'MSE_EKF_dB_avg_partial': MSE_EKF_dB_avg_partial,
# # 'MSE_EKF_linear_arr_partialoptr': MSE_EKF_linear_arr_partialoptr,
# # 'MSE_EKF_dB_avg_partialoptr': MSE_EKF_dB_avg_partialoptr,
# 'KNet_MSE_test_linear_arr': KNet_MSE_test_linear_arr,
# 'KNet_MSE_test_dB_avg': KNet_MSE_test_dB_avg,
# }, EKFfolderName+EKFResultName)