-
Notifications
You must be signed in to change notification settings - Fork 34
/
Copy pathcomp_NB.m
executable file
·280 lines (193 loc) · 8.09 KB
/
comp_NB.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
% simulation code for fast hybrid precoding algorithm for OFDM.
% Created by H.Kasai on July 01, 2018
% Modified from the original codes in
% https://github.com/yuxianghao/Alternating-minimization-algorithms-for-hybrid-precoding-in-millimeter-wave-MIMO-systems
clear
clc
close all;
%% set parameters
measure_snr_flag = true;
realization = 100;
algorithms = { 'Optimal', 'OMP', 'PE-AltMin', 'FPS-AltMin', 'Proposed'};
alg_num = length(algorithms);
%Nt = 144; % # of transmit antennas
%Nr = 36; % # of receive antennas
Nt = 16*16; % # of transmit antennas
Nr = 8*8; % # of receive antennas
Nc = 5; % # of clusters
Nray = 10; % # of rays in each cluster
angle_sigma = 10/180*pi; %standard deviation of the angles in azimuth and elevation both of Rx and Tx
gamma = sqrt((Nt*Nr)/(Nc*Nray)); %normalization factor
sigma = 1; %according to the normalization condition of the H
if measure_snr_flag
Ns = 4;
NRF_array = 6;
%Ns = 8;
%NRF_array = 12;
SNR_dB = -15:5:10;
else
Ns = 6;
NRF_array = 6:1:11;
SNR_dB = 0;
end
SNR = 10.^(SNR_dB./10);
snr_num = length(SNR);
nrf_num = length(NRF_array);
%% set params
clear params;
params.Ns = Ns;
params.Nc = Nc;
params.Nray = Nray;
params.Nt = Nt;
params.Nr = Nr;
params.angle_sigma = angle_sigma;
params.gamma = gamma;
params.sigma = sigma;
params.realization = realization;
%% assign and initialize cells
alg_elapsed_time = cell(alg_num, nrf_num);
info_cell = cell(alg_num, realization);
speceff = cell(alg_num, nrf_num);
speceff_sd = cell(alg_num, nrf_num);
for j=1:alg_num
for k=1:nrf_num
alg_elapsed_time{j,k} = 0;
end
end
c = 1/sqrt(Nc)*exp(1i*[0:2*pi/Nc:2*pi-2*pi/Nc])';
%% perform algorithms
for nrf_idx = 1:length(NRF_array)
NRF = NRF_array(nrf_idx);
params.NRF = NRF;
C = kron(eye(NRF),c);
fprintf('##################### NRF = %d ##################### \n', NRF);
RM = zeros(alg_num, snr_num, realization);
for reali = 1:realization
fprintf('# %d/%d \n', reali, realization);
%% generate matrices for OFDM
[Fopt, Wopt, H, At, Ar, FRF_enc, FRF_dec] = generate_NB_matrices(params);
%% perform actual algorithms
for alg_idx = 1 : alg_num
alg_name = algorithms{alg_idx};
start_time = tic();
switch alg_name
case 'Optimal'
% Do nothing
case 'MO-AltMin'
[FRFM, FBBM, info_cell{alg_idx,reali}] = MO_AltMin_NB_HK(Fopt, NRF, FRF_enc);
FBBM = sqrt(Ns) * FBBM / norm(FRFM * FBBM, 'fro');
[WRFM, WBBM, ~] = MO_AltMin_NB_HK( Wopt, NRF, FRF_dec );
case 'Proposed'
%options_pro.solver = 'tr';
options_pro.func_tolerance = 1e-2;
[FRFM, FBBM, info_cell{alg_idx,reali} ] = fast_hybrid_precoding_NB_wrapper( Fopt, NRF, FRF_enc, options_pro);
FBBM = sqrt(Ns) * FBBM / norm(FRFM * FBBM, 'fro');
[WRFM, WBBM, ~] = fast_hybrid_precoding_NB_wrapper( Wopt, NRF, FRF_dec, options_pro);
case 'PE-AltMin'
[FRF, FBB, info_cell{alg_idx,reali}] = PE_AltMin_NB_HK(Fopt, NRF, FRF_enc);
FBB = sqrt(Ns) * FBB / norm(FRF * FBB, 'fro');
[WRF, WBB, ~] = PE_AltMin_NB_HK( Wopt, NRF, FRF_dec );
case 'OMP'
[FRFO, FBBO, info_cell{alg_idx,reali}] = OMP_NB_HK(Fopt, NRF, At);
FBBO = sqrt(Ns) * FBBO / norm(FRFO * FBBO, 'fro');
[WRFO, WBBO, ~] = OMP_NB_HK(Wopt, NRF, Ar);
case 'FPS-AltMin'
[FRF, FBB] = my_AltMin_new_HK(Fopt, C);
FBB = sqrt(Ns) * FBB / norm(FRF * FBB, 'fro');
[WRF, WBB] = my_AltMin_new_HK(Wopt, C);
end
elapsed_time = toc(start_time);
alg_elapsed_time{alg_idx, nrf_idx} = alg_elapsed_time{alg_idx, nrf_idx} + elapsed_time;
%% Calculate spectral efficiency
for s = 1:snr_num
if strcmp(alg_name, 'MO-AltMin') || strcmp(alg_name, 'Proposed')
RM(alg_idx,s,reali) = RM(alg_idx,s,reali) + log2(det(eye(Ns) + SNR(s)/Ns * pinv(WRFM * WBBM) * H * FRFM * (FBBM * FBBM') * FRFM' * H' * WRFM * WBBM));
elseif strcmp(alg_name, 'PE-AltMin')
RM(alg_idx,s,reali) = RM(alg_idx,s,reali) + log2(det(eye(Ns) + SNR(s)/Ns * pinv(WRF * WBB) * H * FRF * (FBB * FBB') * FRF' * H' * WRF * WBB));
elseif strcmp(alg_name, 'OMP')
RM(alg_idx,s,reali) = RM(alg_idx,s,reali) + log2(det(eye(Ns) + SNR(s)/Ns * pinv(WRFO * WBBO) * H * FRFO * (FBBO * FBBO') * FRFO' * H' * WRFO * WBBO));
elseif strcmp(alg_name, 'FPS-AltMin')
RM(alg_idx,s,reali) = RM(alg_idx,s,reali) + log2(det(eye(Ns) + SNR(s)/Ns * pinv(WRF*WBB) * H * FRF * (FBB * FBB') * FRF' * H' * WRF*WBB));
elseif strcmp(alg_name, 'Optimal')
RM(alg_idx,s,reali) = RM(alg_idx,s,reali) + log2(det(eye(Ns) + SNR(s)/Ns * pinv(Wopt) * H * (Fopt * Fopt') * H' * Wopt));
end
end
fprintf('[%s] \t%04d: %5.2f\n', alg_name, reali, real(RM(alg_idx,snr_num,reali)));
end % end of algorithms
end % end of realizations
fprintf('\n');
for alg_idx = 1 : alg_num
alg_name = algorithms{alg_idx};
speceff{alg_idx, nrf_idx} = sum(real(RM(alg_idx,:,:)),3)/realization;
speceff_sd{alg_idx, nrf_idx} = std(real(RM(alg_idx,:,:)),0,3);
snr = speceff{alg_idx, nrf_idx};
for s = 1:snr_num
fprintf('## [%s] %d[dB]: \t%5.2f, %5.2f [sec]\n', alg_name, SNR_dB(s), snr(s), alg_elapsed_time{alg_idx, nrf_idx});
end
end
end
%% plotting
fs = 20;
linewidth = 2;
style = {'-.+','-d','-*','-s','-o'};
line_color = {[255, 128, 0], [76, 153, 0], [0, 0, 255], [0, 255, 255], [255, 0, 0]};
% generate legend string
legend_str = cell(length(algorithms),1);
for alg_idx = 1 : length(algorithms)
legend_str{alg_idx} = algorithms{alg_idx};
end
if measure_snr_flag
if snr_num == 1
return;
end
% Spectral efficiency for each NRF
for nrf_idx = 1:length(NRF_array)
NRF = NRF_array(nrf_idx);
title_str = sprintf('Spectral Efficiency for SNR (NRF=%d)', NRF);
figure
title(title_str)
for alg_idx = 1 : alg_num
se = zeros(1,nrf_num);
snr = speceff{alg_idx, nrf_idx};
plot(SNR_dB, snr, style{alg_idx}, 'LineWidth', linewidth, 'Color', line_color{alg_idx}/255)
hold on
end
grid on
hold off
ax1 = gca;
set(ax1,'FontSize',fs);
xlabel('SNR [dB]')
ylabel('Spectral efficiency (bits/s/Hz)')
legend(legend_str)
title(title_str)
end
else
if nrf_num == 1
return;
end
% Spectral efficiency when NRF changes
for s=1:snr_num
title_str = sprintf('Spectral Efficiency when NRF changes (%d[dB])', SNR_dB(s));
figure
title(title_str)
for alg_idx = 1 : alg_num
se = zeros(1,nrf_num);
for nrf_idx = 1:nrf_num
snr = speceff{alg_idx, nrf_idx};
snr_for_s(nrf_idx) = snr(s);
%snr_sd = speceff_sd{alg_idx, nrf_idx};
%snrsd_for_s(nrf_idx) = snr_sd(s);
end
plot(NRF_array, snr_for_s, style{alg_idx}, 'LineWidth', linewidth, 'Color', line_color{alg_idx}/255)
hold on
end
grid on
hold off
ax1 = gca;
set(ax1,'FontSize',fs);
xlabel('SNR [dB]')
ylabel('Spectral efficiency (bits/s/Hz)')
legend(legend_str)
title(title_str)
end
end