-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathCNNcontroller.m
181 lines (175 loc) · 8.06 KB
/
CNNcontroller.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
classdef CMACcontroller
properties
in_max;
in_min;
resolusion;
c;
weight;
in_dimension;
out_dimension;
NAP;
PrimeList=[ 2, 3, 5, 7, 11, 13, 17, 19, 23, 29, 31, 37, 41, ...
43, 47, 53, 59, 61, 67, 71, 73, 79, 83, 89, 97, 113, ...
193, 241, 257, 337, 353, 401, 433, 449, 577, 593, 641, ...
673, 769, 881, 929, 977, 1009, 1153, 1201, 1217, 1249, ...
1297,1361, 1409, 1489, 1553, 1601, 1697, 1777, 1873, ...
1889, 2017, 2081, 2113, 2129, 2161, 2273, 2417, 2593, ...
2609, 2657, 2689, 2753, 2801, 2833, 2897, 3041, 3089, ...
3121, 3137, 3169, 3217, 3313, 3329, 3361, 3457, 3617, ...
3697, 3761, 3793, 3889, 4001, 4049, 4129, 4177, 4241, ...
4273, 4289, 4337, 4481, 4513, 4561, 4657, 4673, 4721, ...
4801, 4817, 4993, 5009, 5153, 5233, 5281, 5297, 5393, ...
5441, 5521, 5569, 5857, 5953, 6113, 6257, 6337, 6353, ...
6449, 6481, 6529, 6577, 6673, 6689, 6737, 6833, 6961, ...
6977, 7057, 7121, 7297, 7393, 7457, 7489, 7537, 7649, ...
7681, 7793, 7841, 7873, 7937, 8017, 8081, 8161, 8209, ...
8273, 8353, 8369, 8513, 8609, 8641, 8689, 8737, 8753, ...
8849, 8929, 9041, 9137, 9281, 9377, 9473, 9521, 9601, ...
9649, 9697, 9857];
end
methods
function obj=CMACcontroller(limit, c, out_dimension)%limit[[in_max in_min]'*in_dimension], resolution[1*in_dimension], c[1*1]
obj.in_dimension=length(limit(1,:));
obj.out_dimension=out_dimension;
for i=1:obj.in_dimension
obj.in_max(i)=limit(1,i);
obj.in_min(i)=limit(2,i);
obj.resolusion(i)=(obj.in_max(i)-obj.in_min(i))/(100*c);
end
%NAP=10*c+1;%È¡10c×î½üµÄÖÊÊý
NAP=FindNearestPrime(10*c, obj.PrimeList);
obj.weight=WeightInit(obj.in_dimension, obj.out_dimension, NAP);
obj.c=c;
obj.NAP=NAP;
% obj.weight=zeros(NAP,NAP,obj.out_dimension);
end
function [obj, out] = CMACrecaller(obj, in, outref, trainning_flag)
alpha=0.5;
M=zeros(obj.in_dimension,obj.c);
for i=1:obj.in_dimension
in_quant(i)=floor((in(i)-obj.in_min(i))/obj.resolusion(i));
end
for i=1:obj.c
for j=1:obj.in_dimension
M(j,mod((in_quant(j)+i),obj.c)+1)=in_quant(j)+i;
end
end
out=zeros(1,obj.out_dimension);
for j=1:obj.out_dimension
for i=1:obj.c
switch obj.in_dimension
case 1
out(j)=obj.weight(mod(M(1,i),obj.NAP)+1, j)+out(j);
case 2
out(j)=obj.weight(mod(M(1,i),obj.NAP)+1, mod(M(2,i),obj.NAP)+1, j)+out(j);
case 3
out(j)=obj.weight(mod(M(1,i),obj.NAP)+1, mod(M(2,i),obj.NAP)+1, mod(M(3,i),obj.NAP)+1, j)+out(j);
case 4
out(j)=obj.weight(mod(M(1,i),obj.NAP)+1, mod(M(2,i),obj.NAP)+1, mod(M(3,i),obj.NAP)+1, mod(M(4,i),obj.NAP)+1, j)+out(j);
case 5
out(j)=obj.weight(mod(M(1,i),obj.NAP)+1, mod(M(2,i),obj.NAP)+1, mod(M(3,i),obj.NAP)+1, mod(M(4,i),obj.NAP)+1, mod(M(5,i),obj.NAP)+1, j)+out(j);
case 6
out(j)=obj.weight(mod(M(1,i),obj.NAP)+1, mod(M(2,i),obj.NAP)+1, mod(M(3,i),obj.NAP)+1, mod(M(4,i),obj.NAP)+1, mod(M(5,i),obj.NAP)+1, mod(M(6,i),obj.NAP)+1, j)+out(j);
case 9
out(j)=obj.weight(mod(M(1,i),obj.NAP)+1, mod(M(2,i),obj.NAP)+1, mod(M(3,i),obj.NAP)+1, mod(M(4,i),obj.NAP)+1, mod(M(5,i),obj.NAP)+1, mod(M(6,i),obj.NAP)+1, mod(M(7,i),obj.NAP)+1, mod(M(8,i),obj.NAP)+1, mod(M(9,i),obj.NAP)+1, j)+out(j);
end
end
end
if trainning_flag
for j=1:obj.out_dimension
trainning_error=outref(j)-out(j);
% trainning_error
% % switch j
% % case 1
% % deadzone=5*1e-4;
% % case 2
% % deadzone=5*1e-4;
% % case 3
% % deadzone=5*1e-4;
% % end
% % if abs(trainning_error)<deadzone
% % trainning_error=0;
% % end
% % if trainning_error>deadzone
% % trainning_error=trainning_error-deadzone;
% % end
% % if trainning_error<-deadzone
% % trainning_error=trainning_error+deadzone;
% % end
% trainning_error
for i=1:obj.c
switch obj.in_dimension
case 1
obj.weight(mod(M(1,i),obj.NAP)+1, j)=...
obj.weight(mod(M(1,i),obj.NAP)+1, j)+alpha*(trainning_error)/obj.c;
case 2
obj.weight(mod(M(1,i),obj.NAP)+1, mod(M(2,i),obj.NAP)+1, j)=...
obj.weight(mod(M(1,i),obj.NAP)+1, mod(M(2,i),obj.NAP)+1, j)+alpha*(trainning_error)/obj.c;
case 3
obj.weight(mod(M(1,i),obj.NAP)+1, mod(M(2,i),obj.NAP)+1, mod(M(3,i),obj.NAP)+1, j)=...
obj.weight(mod(M(1,i),obj.NAP)+1, mod(M(2,i),obj.NAP)+1, mod(M(3,i),obj.NAP)+1, j)+alpha*(trainning_error)/obj.c;
case 4
obj.weight(mod(M(1,i),obj.NAP)+1, mod(M(2,i),obj.NAP)+1, mod(M(3,i),obj.NAP)+1, mod(M(4,i),obj.NAP)+1, j)=...
obj.weight(mod(M(1,i),obj.NAP)+1, mod(M(2,i),obj.NAP)+1, mod(M(3,i),obj.NAP)+1, mod(M(4,i),obj.NAP)+1, j)+alpha*(trainning_error)/obj.c;
case 5
obj.weight(mod(M(1,i),obj.NAP)+1, mod(M(2,i),obj.NAP)+1, mod(M(3,i),obj.NAP)+1, mod(M(4,i),obj.NAP)+1, mod(M(5,i),obj.NAP)+1, j)=...
obj.weight(mod(M(1,i),obj.NAP)+1, mod(M(2,i),obj.NAP)+1, mod(M(3,i),obj.NAP)+1, mod(M(4,i),obj.NAP)+1, mod(M(5,i),obj.NAP)+1, j)+alpha*(trainning_error)/obj.c;
case 6
obj.weight(mod(M(1,i),obj.NAP)+1, mod(M(2,i),obj.NAP)+1, mod(M(3,i),obj.NAP)+1, mod(M(4,i),obj.NAP)+1, mod(M(5,i),obj.NAP)+1, mod(M(6,i),obj.NAP)+1, j)=...
obj.weight(mod(M(1,i),obj.NAP)+1, mod(M(2,i),obj.NAP)+1, mod(M(3,i),obj.NAP)+1, mod(M(4,i),obj.NAP)+1, mod(M(5,i),obj.NAP)+1, mod(M(6,i),obj.NAP)+1, j)+alpha*(trainning_error)/obj.c;
case 9
obj.weight(mod(M(1,i),obj.NAP)+1, mod(M(2,i),obj.NAP)+1, mod(M(3,i),obj.NAP)+1, mod(M(4,i),obj.NAP)+1, mod(M(5,i),obj.NAP)+1, mod(M(6,i),obj.NAP)+1, mod(M(7,i),obj.NAP)+1, mod(M(8,i),obj.NAP)+1, mod(M(9,i),obj.NAP)+1, j)=...
obj.weight(mod(M(1,i),obj.NAP)+1, mod(M(2,i),obj.NAP)+1, mod(M(3,i),obj.NAP)+1, mod(M(4,i),obj.NAP)+1, mod(M(5,i),obj.NAP)+1, mod(M(6,i),obj.NAP)+1, mod(M(7,i),obj.NAP)+1, mod(M(8,i),obj.NAP)+1, mod(M(9,i),obj.NAP)+1, j)+alpha*(trainning_error)/obj.c;
end
end
end
%outref-out
end
end
end
end
function weight=WeightInit(in_dimension, out_dimension, NAP)
switch in_dimension
case 1
weight=zeros(NAP, out_dimension);
case 2
weight=zeros(NAP, NAP, out_dimension);
case 3
weight=zeros(NAP, NAP, NAP, out_dimension);
case 4
weight=zeros(NAP, NAP, NAP, NAP, out_dimension);
case 5
weight=zeros(NAP, NAP, NAP, NAP, NAP, out_dimension);
case 6
weight=zeros(NAP, NAP, NAP, NAP, NAP, NAP, out_dimension);
case 9
weight=zeros(NAP, NAP, NAP, NAP, NAP, NAP, NAP, NAP, NAP, out_dimension);
end
end
function y=FindNearestPrime(x, PrimeList)
% while 1==1
% if IsPrime(x, PrimeList)
% y=x;
% break
% end
% x=x+1;
% end
for i=1:length(PrimeList)
if PrimeList(i)>x
y=PrimeList(i);
break
end
end
end
function primeflag=IsPrime(x, PrimeList)
for i=1:length(PrimeList)
if length(PrimeList)/2+1<i
primeflag=true;
break
end
if mod(x,PrimeList(i))==0
primeflag=false;
break
end
end
end