提交 ba30d623 编写于 作者: LSayhi's avatar LSayhi

update

上级 66c73cf7
......@@ -2,11 +2,14 @@
Welcome everyone to exchange and correct, if you feel helpful, thanks for star and fork.
みんなを歓迎して交流と指摘、もし助けがあって、ありがとうございますをクリックしてスター.
project 1:基于神经网络的带通滤波器优化设计
project 1:基于深度学习的滤波器优化设计
version 1.0 start time :2017.12.17-
project 2:双频微带天线设计与神经网络优化
version 1.0 start time : 2017.12.21-
project 3:Coursera-deep learning.ai
version 1.0 start time : 2018.1.16-
\ No newline at end of file
project 3:Coursera:deep learning深度学习
version 1.0 start time : 2018.1.16-
project 3:宽频带范围螺旋电感DNN建模
version 1.0 start time : 2018.5.8-
\ No newline at end of file
名称:基于神经网络的带通滤波器设计优化
名称:基于深度学习的滤波器设计优化
内容:
利用HFSS在32-36GHz的条件下,仿真了一个带通滤波器结构;
利用仿真的数据数据S1=0.01:0.04:0.21,S2=0.26:0.02:0.46,W1=0.01:0.04:0.21,Freq=32:36
......@@ -9,5 +9,5 @@
说明:
neuralnetwork.m 文件是matlab程序
训练集测试集数据 文件夹是HFSS导出的原始数据和matlab处理后的数据
基于神经网络的带通滤波器设计仿真源文件.aedt 文件是HFSS仿真文件
基于神经网络的带通滤波器设计优化演讲版.pptx 文件是此项目的说明ppt
\ No newline at end of file
基于深度学习的滤波器设计仿真源文件.aedt 文件是HFSS仿真文件
基于深度学习的滤波器设计优化说明.pptx 文件是此项目的说明ppt
\ No newline at end of file
%DNN电感建模
%Author:Liujiangfan
%Time :2018.5
%Github:https://github.com/LSayhi
%Addres:SCHOOL OF ELECTRONIC ENGINEERING at BUPT,Beijing ,China
%说明
%输入数据依次为频率(标准化/60G),电感内径(/29.6um),电感线宽(/5.1um)
%输出数据依次为S11模值,S21模值,S11相位(弧度制),S21相位(弧度制)
%输入输出若不标准化则可能导致模型不收敛或收敛变慢
%定义
num_inputlayer=3;%输入参数个数
num_outputlayer=4;%输出参数个数
input_train = traininput(1:num_inputlayer,:);%读入训练集输入
output_train= trainoutput(1:num_outputlayer,:);%读入训练集输出
input_test = testinput(1:num_inputlayer,:);%读入测试集输入
output_test = testoutput(1:num_outputlayer,:);%读入测试集输出
[inputn,inputps]=mapminmax(input_train);%归一化训练集输入
[outputn,outputps]=mapminmax(output_train);%归一化训练集输出
net=newff(inputn,outputn,[10,10]);
%{'tansig','tansig','purelin'},激活函数
%设置隐藏层层数和每层神经元数目,输入输出层由数据确定。
net.trainParam.epochs=10000;%最大迭代次数1000
net.trainParam.lr=0.01;%学习率0.1
net.trainParam.goal=0.000001;%目标误差率
% net.divideParam.trainRatio = 60/100;
% net.divideParam.valRatio = 30/100;
% net.divideParam.testRatio = 10/100;
%net.divideFcn = '';
net=train(net,inputn,outputn);%训练设置好的神经网络模型
inputn_test=mapminmax('apply',input_test,inputps);
%测试集输入归一化(按训练集的归一化参数来,不是测试集按自己数据归一化)
an=sim(net,inputn_test);%将测试集输入导入模型net仿真
BPoutput=mapminmax('reverse',an,outputps);
%反归一化测试集输出(按训练集的参数反归一化)
%画直角坐标图
figure(1);
plot(testinput(1,:),BPoutput([1,3],:),'.')%预测S11的幅值和相位
hold on;
plot(testinput(1,:),output_test([1,3],:),'-')%仿真的S11幅值和相位
legend('预测S11mag','预测S11deg','仿真S11mag','仿真S11deg')
title('Simulated S by HFSS and predicted S by neural network')
figure(2);
plot(testinput(1,:),BPoutput([2,4],:),'.')%预测S21的幅值和相位
hold on;
plot(testinput(1,:),output_test([2,4],:),'-')%仿真的S21幅值和相位
legend('预测S21mag','预测S21deg','仿真S21mag','仿真S21deg')
title('Simulated S by HFSS and predicted S by neural network')
%画极坐标图
figure(3);
polar(BPoutput(1,:),BPoutput(3,:),'.');%S11仿真
hold on;
polar(output_test(1,:),output_test(3,:),'-');%S11预测
legend('仿真S11','预测S11')
title('Simulated S by HFSS and predicted S by neural network')
figure(4);
polar(BPoutput(2,:),BPoutput(4,:),'.');%S21仿真
hold on;
polar(output_test(2,:),output_test(4,:),'-');%S21预测
legend('仿真S21','预测S21')
title('Simulated S by HFSS and predicted S by neural network')
名称:宽频带范围螺旋电感DNN建模
内容:
利用HFSS在1-119GHz的条件下,抽样仿真了一个螺旋电感,利用仿真数据建立电感模型;
输入为频率(标准化/60GHz)、电感内径(标准化/29.6um)、电感线宽(标准化/5.1um)
输出为MAGS11,MAGS21,DEGS11(弧度制),DEGS21(弧度制)
说明:
model.m 文件是matlab程序
训练集测试集数据 文件夹是HFSS导出的原始数据和matlab预处理后的数据
\ No newline at end of file
"Freq [GHz]","ang_deg(S(1,1)) [deg]"
1,43.8433383195517
2,57.3672016208201
3,64.4186678062062
4,68.140203285827
5,70.1478656809233
6,71.20113621756
7,71.6820392497097
8,71.7987410140467
9,71.6726417727326
10,71.3785036286038
11,70.9644188693213
12,70.4624442434995
13,69.8945973793806
14,69.2764013699421
15,68.6190624083136
16,67.9308503671021
17,67.2179969432527
18,66.4852926616546
19,65.7364910692131
20,64.9745868533481
21,64.2020100471132
22,63.4207635245785
23,62.6325216503176
24,61.8387019906816
25,61.0405181282741
26,60.23901907355
27,59.4351190671019
28,58.6296204183823
29,57.8232312438159
30,57.016579428388
31,56.2102237605114
32,55.4046629278058
33,54.6003428762311
34,53.7976629031313
35,52.9969807600448
36,52.1986169725666
37,51.4028585345009
38,50.6099620966922
39,49.8201567435694
40,49.0336464299622
41,48.2506121352825
42,47.4712137803924
43,46.6955919434393
44,45.9238694039335
45,45.1561525388835
46,44.392532590491
47,43.6330868214989
48,42.8778795715483
49,42.1269632256913
50,41.3803791044146
51,40.6381582830522
52,39.9003223472399
53,39.1668840900406
54,38.4378481554957
55,37.7132116326074
56,36.992964603095
57,36.2770906456788
58,35.5655672991001
59,34.8583664855748
60,34.155454895879
61,33.4567943367652
62,32.7623420408978
63,32.0720509389518
64,31.3858698929366
65,30.703743889174
66,30.0256141886626
67,29.3514184318046
68,28.6810906936616
69,28.0145614850638
70,27.3517576940921
71,26.6926024617809
72,26.0370149855396
73,25.3849102440952
74,24.7361986392053
75,24.0907855528256
76,23.4485708250839
77,22.8094481702388
78,22.1733045675817
79,21.5400196959324
80,20.909465529218
81,20.2815062829755
82,19.655999003805
83,19.0327952292365
84,18.4117443081981
85,17.792699135293
86,17.1755251490091
87,16.5601133468511
88,15.9463975772794
89,15.3343752337038
90,14.7241285381207
91,14.1158410523972
92,13.5098017837775
93,12.9063889663571
94,12.3060292003537
95,11.7091355922027
96,11.1160382352472
97,10.5269264088559
98,9.941819561983
99,9.36057373156551
100,8.78291725041621
101,8.20850130194947
102,7.63695011734991
103,7.06790050089684
104,6.50102684834685
105,5.93605277708914
106,5.37275289848256
107,4.81094862137801
108,4.25050114387636
109,3.69130375180442
110,3.13327462346479
111,2.57635068420195
112,2.02048265194578
113,1.46563120196598
114,0.911764087797286
115,0.358854033160655
116,-0.193122778036665
117,-0.74418776045617
118,-1.2943606527439
119,-1.84366000090274
"Freq [GHz]","mag(S(1,1)) []"
1,0.0232842867812479
2,0.0344760918479893
3,0.0469027204850789
4,0.0599621274501286
5,0.0732885878369999
6,0.0867092234084981
7,0.100135763728635
8,0.113519599223324
9,0.126832626226254
10,0.140058161848076
11,0.15318616014274
12,0.166210479308483
13,0.179127237506631
14,0.19193380062108
15,0.204628160592249
16,0.217208563687633
17,0.229673300530543
18,0.242020600090868
19,0.254248589002555
20,0.266355290330459
21,0.278338644627849
22,0.290196542122677
23,0.301926858972515
24,0.313527493298081
25,0.324996398543759
26,0.336331612906876
27,0.347531284330928
28,0.358593691016634
29,0.369517257668618
30,0.380300567834553
31,0.39094237275401
32,0.401441597146745
33,0.411797342355513
34,0.422008887228907
35,0.432075687093721
36,0.441997371128476
37,0.451773738412938
38,0.461404752894063
39,0.470890537477588
40,0.480231367426584
41,0.489427663223557
42,0.498479983031072
43,0.507389014866765
44,0.516155568592159
45,0.52478056780009
46,0.533265041672949
47,0.541610116872838
48,0.549817009515053
49,0.557887017267766
50,0.565821511613343
51,0.573621930300166
52,0.581289770008088
53,0.588826579245649
54,0.596233951492806
55,0.603513518599224
56,0.610666944444968
57,0.617695918867906
58,0.624602151860077
59,0.631387368033924
60,0.638053301358605
61,0.644601690166689
62,0.651034272432693
63,0.657352781327207
64,0.663558941054333
65,0.669654462986138
66,0.67564104211662
67,0.681520353870131
68,0.687294051316616
69,0.6929637628701
70,0.698531090579775
71,0.703997609167881
72,0.709364866029083
73,0.714634382487041
74,0.719807656711067
75,0.724886168835691
76,0.729871389004604
77,0.734764789281866
78,0.739567860634713
79,0.744282136477008
80,0.748909224525515
81,0.753450848867772
82,0.757908903991251
83,0.762285521770678
84,0.766583150570479
85,0.770804642018592
86,0.774953334874511
87,0.779033116217619
88,0.783048428406021
89,0.787004178710735
90,0.790905503841266
91,0.794757354471575
92,0.798563906599028
93,0.802327879194316
94,0.806049920977099
95,0.809728276916406
96,0.813358905840322
97,0.816936082185514
98,0.820453334667234
99,0.823904452046069
100,0.827284288955151
101,0.830589217453267
102,0.833817215204996
103,0.836967684720744
104,0.84004113252327
105,0.843038818308178
106,0.845962443228686
107,0.8488139073807
108,0.851595139866995
109,0.85430799103201
110,0.856954171902346
111,0.859535226438162
112,0.862052524899471
113,0.86450726973619
114,0.866900508133665
115,0.869233147443177
116,0.871505971218589
117,0.873719654580844
118,0.875874778271358
119,0.877971841144947
"Freq [GHz]","ang_deg(S(2,1)) [deg]"
1,-1.18899865813171
2,-2.20546116935948
3,-3.24945603470357
4,-4.2999843693872
5,-5.35040957657264
6,-6.39799970750373
7,-7.44142495558406
8,-8.48001141970906
9,-9.51344453608911
10,-10.5416188049201
11,-11.5645480416801
12,-12.5823074442162
13,-13.5949957451526
14,-14.6027112844981
15,-15.605537954999
16,-16.6035380360765
17,-17.5967496559717
18,-18.5851871986748
19,-19.5688434435338
20,-20.5476925993795
21,-21.5216936792817
22,-22.4907938691134
23,-23.4549316879323
24,-24.4140398355056
25,-25.3680476850608
26,-26.3168834178061
27,-27.2604758178434
28,-28.1987557574306
29,-29.1316574071439
30,-30.0591192060216
31,-30.9810846250327
32,-31.8975027543147
33,-32.8083287412974
34,-33.7135241034503
35,-34.613056936217
36,-35.5069020338252
37,-36.3950409381337
38,-37.2774619284941
39,-38.1541599637539
40,-39.0251365859652
41,-39.8903997940641
42,-40.7499638947193
43,-41.6038493366674
44,-42.452082534155
45,-43.2946956845444
46,-44.1317265847041
47,-44.9632184504737
48,-45.7892197432389
49,-46.6097840074566
50,-47.4249697227969
51,-48.2348401743682
52,-49.0394633441919
53,-49.8389118265796
54,-50.6332627691524
55,-51.4225978396451
56,-52.2070032158918
57,-52.986569591778
58,-53.7613921843211
59,-54.5315707146228
60,-55.2972093154624
61,-56.0584162865249
62,-56.8153035683518
63,-57.5679857288486
64,-58.3165781390175
65,-59.0611938420828
66,-59.8019383779567
67,-60.5389015117824
68,-61.2721444730176
69,-62.0016810898352
70,-62.7274514774441
71,-63.4492884716934
72,-64.1668810288328
73,-64.8797466378807
74,-65.5872360127892
75,-66.2886021333938
76,-66.98315769367
77,-67.6705038411832
78,-68.3507456873629
79,-69.0245699295901
80,-69.6931117693912
81,-70.3576663043905
82,-71.0193921396065
83,-71.6791314815785
84,-72.3373746661724
85,-72.9943217845321
86,-73.6499781259376
87,-74.3042412977988
88,-74.9569638812238
89,-75.6079916247892
90,-76.2571833117468
91,-76.9044189341455
92,-77.5496012609954
93,-78.1926541166512
94,-78.8335193124051
95,-79.472153270644
96,-80.1085238418848
97,-80.7426075164557
98,-81.3743870787453
99,-82.0038496797756
100,-82.6309852752872
101,-83.2557853698738
102,-83.8782420111108
103,-84.4983469849492
104,-85.1160911718227
105,-85.7314640305948
106,-86.3444531841059
107,-86.9550440855874
108,-87.5632197496476
109,-88.1689605350703
110,-88.7722439694427
111,-89.3730446078082
112,-89.9713339192348
113,-90.5670801965126
114,-91.1602484852232
115,-91.7508005292252
116,-92.3386947302239
117,-92.9238861195837
118,-93.5063263409203
119,-94.0859636423075
"Freq [GHz]","mag(S(2,1)) []"
1,0.982990683567484
2,0.981929167371244
3,0.981101321002651
4,0.980103543394259
5,0.978888725338347
6,0.977456939966372
7,0.975818751061639
8,0.973987031762796
9,0.971974313772976
10,0.969791793641815
11,0.967449025685557
12,0.964953951658647
13,0.96231309592875
14,0.959531827694784
15,0.956614632240654
16,0.95356535932083
17,0.950387433888091
18,0.947084025039252
19,0.943658175123702
20,0.940112893933831
21,0.936451223937328
22,0.93267628245024
23,0.928791286038022
24,0.924799561608189
25,0.920704547813619
26,0.916509789615867
27,0.912218928201563
28,0.907835687908633
29,0.903363861393705
30,0.89880729394171
31,0.894169867566263
32,0.889455485359062
33,0.884668056404174
34,0.879811481467796
35,0.874889639596436
36,0.869906375699807
37,0.86486548915339
38,0.859770723425787
39,0.85462575671435
40,0.849434193557006
41,0.844199557376926
42,0.838925283908294
43,0.833614715445185
44,0.828271095850298
45,0.822897566255723
46,0.817497161383125
47,0.812072806405425
48,0.806627314265361
49,0.801163383357777
50,0.795683595471045
51,0.79019041386783
52,0.78468618136494
53,0.77917311824466
54,0.773653319793523
55,0.768128753216178
56,0.762601253608507
57,0.757072518591705
58,0.75154410110368
59,0.746017399713099
60,0.740493645665482
61,0.734973885699208
62,0.729458959510532
63,0.723949470667132
64,0.718445749909243
65,0.712947810411393
66,0.707455296216507
67,0.701967428587252
68,0.696482961869677
69,0.691000172597745
70,0.685516924908333
71,0.680030881558663
72,0.674539954883285
73,0.669043090479205
74,0.66354139732513
75,0.658039421100932
76,0.652546012580484
77,0.647073985044225
78,0.641638005984488
79,0.636251140382176
80,0.630921563355391
81,0.625651077427646
82,0.620435918897783
83,0.615268986368249
84,0.610142190047817
85,0.605048093341374
86,0.599980689613958
87,0.594935543000544
88,0.589909594378602
89,0.584900853996256
90,0.579908100640865
91,0.57493063486233
92,0.569968094738409
93,0.56502032622805
94,0.560087295686681
95,0.555169032819219
96,0.550265594701751
97,0.545377043974216
98,0.540503436359827
99,0.53564481420832
100,0.530801203855152
101,0.525972615341892
102,0.521159043550261
103,0.516360470139367
104,0.511576865897458
105,0.506808193264349
106,0.502054408874629
107,0.497315466032242
108,0.492591317065629
109,0.487881915536954
110,0.483187218294018
111,0.478507187362583
112,0.473841791682089
113,0.469191008690617
114,0.46455482576628
115,0.459933241532645
116,0.455326267035637
117,0.450733926798966
118,0.446156259764533
119,0.44159332012364
Markdown is supported
0% .
You are about to add 0 people to the discussion. Proceed with caution.
先完成此消息的编辑!
想要评论请 注册