/
FigTrainingBehavior.m
159 lines (131 loc) · 5.02 KB
/
FigTrainingBehavior.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
% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
% Routine for obtaining the behavior of the adaptive algorithms as they
% progress. Figure 4.
%
% Reference:
%
% [1] Ramirez, J., & Paredes, J. (2016). Recursive Weighted Myriad Based
% Filters and their Optimizations. IEEE Transactions on Signal
% Processing, 64(15), 4027-4039.
%
% Author:
% Juan Marcos Ramirez, M.S.
% Universidad de Los Andes, Merida, Venezuela
% email: juanra@ula.ve, juanmarcos26@gmail.com
%
% Date:
% September, 2016
% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
clear all;
close all;
disp('---------------------------------------');
disp('This routine could take several minutes');
disp('---------------------------------------');
% Routine for obtaining the behavior of the adaptive algorithms
addpath('RecursiveMyriadFilters/');
% Adaptive recursive myriad based filter parameters
Ki = 1;
u = 0.001;
M1 = 64;
M2 = 32;
% Training signal parameters
N = 5000;
% Simulation parameters
trials = 10;
% FIR filter parameters
FILTER_ORDER = 96;
firFilterDesign = fir1(FILTER_ORDER - 1,[.075 .15]);
maeRWMyEnsembleAverage = zeros(N - M1 + 1,1);
maeRHMyEnsembleAverage = zeros(N - M1 + 1,1);
maeRMMyEnsembleAverage = zeros(N - M1 + 1,1);
maeSRWMyEnsembleAverage = zeros(N - M1 + 1,1);
maeSRHMyEnsembleAverage = zeros(N - M1 + 1,1);
gSRHMyEnsembleAverage = zeros(N - M1 + 1,M1);
hSRHMyEnsembleAverage = zeros(N - M1 + 1,M2);
KSRHMyEnsembleAverage = zeros(1,N - M1 + 1);
gSRWMyEnsembleAverage = zeros(N - M1 + 1,M1);
hSRWMyEnsembleAverage = zeros(N - M1 + 1,M2);
K1SRWMyEnsembleAverage = zeros(1,N - M1 + 1);
K2SRWMyEnsembleAverage = zeros(1,N - M1 + 1);
disp(['Number of realizations: ' num2str(trials)])
for ii = 1:trials
tic;
disp(['Iteration: ' num2str(ii)]);
% Training signal settings
tempTrainingSignal = sign(randn(1,N + 48));
trainingSignal = tempTrainingSignal(1:N);
tempDesiredSignal = filter(firFilterDesign,1,tempTrainingSignal);
desiredSignal = tempDesiredSignal(48:end);
h = ones(1,M2)/(M1 + M2);
g = ones(1,M1)/(M1 + M2);
[~, ~, eRWMy] = adaptiveRWMy(trainingSignal,desiredSignal, g, h, Ki, Ki, u);
[~, ~, eRHMy] = adaptiveRHMy(trainingSignal,desiredSignal, g, h, Ki, u);
maeRWMyEnsembleAverage = maeRWMyEnsembleAverage + abs(eRWMy)/trials;
maeRHMyEnsembleAverage = maeRHMyEnsembleAverage + abs(eRHMy)/trials;
[~, ~, ~, ~, eSRWMy, gSRWMy, hSRWMy, K1SRWMy, K2SRWMy] = adaptiveSRWMy(trainingSignal,desiredSignal, g, h, Ki, Ki, u);
[~, ~, ~, eSRHMy, gSRHMy, hSRHMy, KSRHMy] = adaptiveSRHMy(trainingSignal,desiredSignal, g, h, Ki, u);
maeSRWMyEnsembleAverage = maeSRWMyEnsembleAverage + abs(eSRWMy)/trials;
maeSRHMyEnsembleAverage = maeSRHMyEnsembleAverage + abs(eSRHMy)/trials;
gSRHMyEnsembleAverage = gSRHMyEnsembleAverage + gSRHMy/trials;
hSRHMyEnsembleAverage = hSRHMyEnsembleAverage + hSRHMy/trials;
KSRHMyEnsembleAverage = KSRHMyEnsembleAverage + KSRHMy/trials;
gSRWMyEnsembleAverage = gSRWMyEnsembleAverage + gSRWMy/trials;
hSRWMyEnsembleAverage = hSRWMyEnsembleAverage + hSRWMy/trials;
K1SRWMyEnsembleAverage = K1SRWMyEnsembleAverage + K1SRWMy/trials;
K2SRWMyEnsembleAverage = K2SRWMyEnsembleAverage + K2SRWMy/trials;
toc;
end
maeRWMyTrial = abs(eRWMy);
maeRHMyTrial = abs(eRHMy);
maeSRWMyTrial = abs(eSRWMy);
maeSRHMyTrial = abs(eSRHMy);
%% Display Results
n = 1:1:N-M1+1;
n_rand1 = randi([1 M1],1);
n_rand2 = randi([1 M2],1);
subplot(321)
plot(n, gSRWMy(:,n_rand1)), hold on;
plot(n, gSRWMyEnsembleAverage(:,n_rand1), 'k');
plot(n, hSRWMy(:,n_rand2), 'r');
plot(n, hSRWMyEnsembleAverage(:,n_rand2), 'k--');
title('Scaled Recursive Weighted Myriad')
ylabel('Parameter value'), xlabel('Iteration (n)');
legend('g (trial)', 'g (average)', 'h (trial)', 'h (average)');
axis('tight');
subplot(322)
plot(n, K1SRWMy), hold on;
plot(n, K1SRWMyEnsembleAverage, 'k');
plot(n, K2SRWMy, 'r');
plot(n, K2SRWMyEnsembleAverage, 'k--');
title('Scaled Recursive Weighted Myriad')
ylabel('Linearity parameters'), xlabel('Iteration (n)');
legend('K1 (trial)', 'K1 (average)', 'K2 (trial)', 'K2 (average)');
axis('tight');
subplot(323)
plot(n, maeRWMyTrial), hold on;
plot(n, maeRWMyEnsembleAverage, 'k');
legend('Single Trial', 'Ensemble Average');
title('Recursive Weighted Myriad')
ylabel('MAE'), xlabel('Iteration (n)');
axis('tight');
subplot(324)
plot(n, maeRHMyTrial), hold on;
plot(n, maeRHMyEnsembleAverage, 'k');
legend('Single Trial', 'Ensemble Average');
title('Recursive Hybrid Myriad')
ylabel('MAE'), xlabel('Iteration (n)');
axis('tight');
subplot(325)
plot(n, maeSRWMyEnsembleAverage), hold on;
plot(n, maeRWMyEnsembleAverage, 'k');
legend('Scaled', 'Normalized');
title('Recursive Weighted Myriad')
ylabel('MAE'), xlabel('Iteration (n)');
axis('tight');
subplot(326)
plot(n, maeSRHMyEnsembleAverage), hold on;
plot(n, maeRHMyEnsembleAverage, 'k');
legend('Scaled', 'Normalized');
title('Recursive Hybrid Myriad')
ylabel('MAE'), xlabel('Iteration (n)');
axis('tight');