Scribd
Upload
93 views18 pages

SC-NOMA vs TDMA Capacity Comparison

This document compares SC-NOMA and TDMA. It simulates channel models for different users, calculates achievable rates and sum rates for different multiple access schemes like NOMA, TDMA and OMA. It also simulates BER performance of NOMA in AWGN and Rayleigh fading channels. Outage probability and achievable rates are plotted with respect to transmit power for NOMA.

Uploaded by

vanlinh le
Copyright
© All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd
93 views18 pages

SC-NOMA vs TDMA Capacity Comparison

This document compares SC-NOMA and TDMA. It simulates channel models for different users, calculates achievable rates and sum rates for different multiple access schemes like NOMA, TDMA and OMA. It also simulates BER performance of NOMA in AWGN and Rayleigh fading channels. Outage probability and achievable rates are plotted with respect to transmit power for NOMA.

Uploaded by

vanlinh le
Copyright
© All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd

So sánh SC-NOMA và TDMA

clc; clear variables;


close all;

d1 = 10; d2 = 9;
d3 = 4; d4 = 3;

eta = 4;

N = 10^4;

h1 = sqrt(d1^-eta)*(randn(1,N)+1i*randn(1,N))/sqrt(2);
h2 = sqrt(d2^-eta)*(randn(1,N)+1i*randn(1,N))/sqrt(2);
h3 = sqrt(d3^-eta)*(randn(1,N)+1i*randn(1,N))/sqrt(2);
h4 = sqrt(d4^-eta)*(randn(1,N)+1i*randn(1,N))/sqrt(2);

g1 = (abs(h1)).^2;
g2 = (abs(h2)).^2;
g3 = (abs(h3)).^2;
g4 = (abs(h4)).^2;

SNR = 20:2:40;
snr = db2pow(SNR);

a1 = 0.7; a2 = 1-a1;

b1 = 0.7; b2 = (3/4)*(1-b1); b3 = (3/4)*(1-(b1+b2)); b4 = 1


- (b1+b2+b3);
T = 1;
epsilon = (2^T)-1;

for u = 1:length(snr)
% a1 = min(1,epsilon*((1/snr(u)) +
g1)./(g1*(1+epsilon)));
% a2 = 1 - a1;
%
% a1 = min(1,epsilon*((1/snr(u)) +
g1)./(g1*(1+epsilon)));
% a2 = 1 - a1;

%1-2 3-4 (n-n, f-f)


C1(u) = mean(log2(1 + a1.*g1*snr(u)./(a2.*g1*snr(u)+1)));
C2(u) = mean(log2(1 + a2.*g2*snr(u)));
C3(u) = mean(log2(1 + a1.*g3*snr(u)./(a2.*g3*snr(u)+1)));
C4(u) = mean(log2(1 + a2.*g4*snr(u)));

%1-4 2-3 (n-f)


C11(u) = mean(log2(1 + a1.*g1*snr(u)./(a2.*g1*snr(u)
+1)));
C44(u) = mean(log2(1 + a2.*g4*snr(u)));
C22(u) = mean(log2(1 + a1.*g2*snr(u)./(a2.*g2*snr(u)
+1)));
C33(u) = mean(log2(1 + a2.*g3*snr(u)));

%TDMA
t1(u) = 0.25*mean(log2(1 + g1*snr(u)));
t2(u) = 0.25*mean(log2(1 + g2*snr(u)));
t3(u) = 0.25*mean(log2(1 + g3*snr(u)));
t4(u) = 0.25*mean(log2(1 + g4*snr(u)));

% %NOMA
n1(u) = mean(log2(1 +
b1.*g1*snr(u)./((b2+b3+b4).*g1*snr(u)+1)));
n2(u) = mean(log2(1 + b2.*g2*snr(u)./((b3+b4).*g2*snr(u)
+1)));
n3(u) = mean(log2(1 + b3.*g3*snr(u)./(b4.*g3*snr(u)+1)));
n4(u) = mean(log2(1 + b4.*g4*snr(u)));
%
% %
end

R1 = 0.5*(C1+C2+C3+C4);
R11 = 0.5*(C11+C22+C33+C44);
t = (t1+t2+t3+t4);
n = n1+n2+n3+n4;
figure;
plot(SNR,R11,'o-b','linewidth',1.5);hold on; grid on;
plot(SNR,R1,'o-r','linewidth',1.5);
plot(SNR,n,'o-m','linewidth',1.5);
plot(SNR,t,'o-k','linewidth',1.5);
legend('Hybrid NOMA N-F pairing','Hybrid NOMA N-N,F-F
pairing','SC-NOMA','TDMA');
xlabel('SNR (dB)');
ylabel('Sum rate (bps/Hz)');
So sánh công suất của OMA và NOMA

clc; clear variables; close all;

%SNR range
Pt = -114:5:-54; %in dB
pt = db2pow(Pt); %in linear scale

N = 10^4;

d1 = 5; d2 = 3; d3 = 2; %Distance of users
eta = 4; %Path loss exponent

%Rayleigh fading coefficients of both users


h1 = (sqrt(d1^-eta))*(randn(N,1) + 1i*randn(N,1))/sqrt(2);
h2 = (sqrt(d2^-eta))*(randn(N,1) + 1i*randn(N,1))/sqrt(2);
h3 = (sqrt(d3^-eta))*(randn(N,1) + 1i*randn(N,1))/sqrt(2);

%Channel gains
g1 = (abs(h1)).^2;
g2 = (abs(h2)).^2;
g3 = (abs(h3)).^2;

BW = 10^9;
No = -174 + 10*log10(BW);
no = (10^-3)*db2pow(No);

%Power allocation coefficients


a1 = 0.75; a2 = 0.1825; a3 = 1-(a1+a2);

C_noma = zeros(1,length(pt));
C_oma = zeros(1,length(pt));
for u = 1:length(pt)

%NOMA capacity calculation


C_noma_1 = log2(1 + pt(u)*a1.*g1./(pt(u)*a2.*g1 +
pt(u)*a3.*g1 + no)); %User 1
C_noma_2 = log2(1 + pt(u)*a2.*g2./(pt(u)*a3.*g1+no));
%User 2
C_noma_3 = log2(1 + pt(u)*a3.*g2/no);
%User 3
C_noma_sum(u) = mean(C_noma_1 + C_noma_2 + C_noma_3);
%Sum capacity of NOMA

%OMA capacity calculation


C_oma_1 = (1/3)*log2(1 + pt(u)*g1/no); %User 1
C_oma_2 = (1/3)*log2(1 + pt(u)*g2/no); %User 2
C_oma_3 = (1/3)*log2(1 + pt(u)*g3/no); %User 3

C_oma_sum(u) = mean(C_oma_1 + C_oma_2 + C_oma_3); %Sum


capacity of OMA

end
SNR = Pt - No;
figure;
plot(SNR,C_noma_sum,'linewidth',2); hold on; grid on;
plot(SNR,C_oma_sum,'linewidth',2)
xlabel('SNR (dB)');
ylabel('Achievable rate (bps/Hz)');
legend('NOMA','OMA');
title('Capacity of NOMA vs OMA');
ylim([0 max(C_noma_sum)+1]);

Mô phỏng BER của NOMA trong kênh AWGN (Nhiễu Gausse trắng cộng sinh)
clc; clear variables; close all;

SNR = 0:30; %SNR range in dB


snr = db2pow(SNR); %SNR range in linear scale

%Tao chuoi nhi phan co do dai N


N = 10^5;

x1 = randi([0 1],1,N); %Du lieu bit cua user 1


x2 = randi([0 1],1,N); %Du lieu bit cua user 2

%Dieu che BPSK(Pha nhi phan)cho du lieu dc tao o tren


xmod1 = 2*x1 - 1;
xmod2 = 2*x2 - 1;

%Thiet lap trong so cong suat cho cac user


a1 = 0.75;
a2 = 0.25;

%Thuc hien ma hoa chong chat


x = sqrt(a1)*xmod1 + sqrt(a2)*xmod2; %sqrt là Ham khai can

%Them AWGN vào x(Truyen x qua kenh AWGN)


for u = 1:length(snr)
y1 = awgn(x,SNR(u),'measured'); %T/h cua user 1 bi tac
dong boi AWGN
y2 = awgn(x,SNR(u),'measured'); %T/h cua user 2 bi tac
dong boi AWGN

%AT USER 1
%Direct decoding of x from y1
x1_hat = ones(1,N); %just a buffer
x1_hat(y1 < 0) = 0;

%AT USER 2
%Direct decoding of x from y2
x11_est = ones(1,N); %just a buffer
x11_est(y2 < 0) = 0; %Estimate user 1's signal first
x11_est(x11_est == 0) = -1; %Remodulate user 1's signal
%Subtract remodulated x11_est component from y2
rem = y2 - sqrt(a1)*x11_est; % T/h dang o dang d/che
BPSK

%Decode x2 from rem


x2_hat = zeros(1,N);
x2_hat(rem>0) = 1;

%Estimate BER
ber1(u) = biterr(x1,x1_hat)/N;% = so bit loi/ tong so
bit
ber2(u) = biterr(x2,x2_hat)/N;
end

%plot BER curves


semilogy(SNR, ber1, 'linewidth', 1.5); hold on;
semilogy(SNR, ber2, 'linewidth', 1.5); grid on;
legend('User 1 \alpha_1 = 0.75','User 2 \alpha_2 = 0.25');
xlabel('SNR (dB)');
ylabel('BER');
title('BER graph for NOMA in AWGN channel');
Dung lượng và xác suất ngừng hoạt động
clc; clear variables; close all;
N = 10^5;

d1 = 1000; d2 = 500; %Distances of users from base


station (BS)
a1 = 0.75; a2 = 0.25; %Power allocation factors
eta = 4; %Path loss exponent

%Generate rayleigh fading coefficient for both users


h1 = sqrt(d1^-eta)*(randn(1,N)+1i*randn(1,N))/sqrt(2);
h2 = sqrt(d2^-eta)*(randn(1,N)+1i*randn(1,N))/sqrt(2);

g1 = (abs(h1)).^2;
g2 = (abs(h2)).^2;

Pt = 0:2:40; %Transmit power in dBm


pt = (10^-3)*10.^(Pt/10); %Transmit power in linear scale
BW = 10^6; %System bandwidth
No = -174 + 10*log10(BW); %Noise power (dBm)
no = (10^-3)*10.^(No/10); %Noise power (linear scale)

p = length(Pt);
p1 = zeros(1,length(Pt));
p2 = zeros(1,length(Pt));

rate1 = 1; rate2 = 2; %Target rate of users in bps/Hz


for u = 1:p
%Calculate SNRs
gamma_1 = a1*pt(u)*g1./(a2*pt(u)*g1+no);
gamma_12 = a1*pt(u)*g2./(a2*pt(u)*g2+no);
gamma_2 = a2*pt(u)*g2/no;

%Calculate achievable rates


R1 = log2(1+gamma_1);
R12 = log2(1+gamma_12);
R2 = log2(1+gamma_2);

%Find average of achievable rates


R1_av(u) = mean(R1);
R12_av(u) = mean(R12);
R2_av(u) = mean(R2);

%Check for outage


for k = 1:N
if R1(k) < rate1
p1(u) = p1(u)+1;
end
if (R12(k) < rate1)||(R2(k) < rate2)
p2(u) = p2(u)+1;
end
end
end

pout1 = p1/N;
pout2 = p2/N;

figure;
semilogy(Pt, pout1, 'linewidth', 1.5); hold on; grid on;
semilogy(Pt, pout2, 'linewidth', 1.5);
xlabel('Transmit power (dBm)');
ylabel('Outage probability');
legend('User 1 (far user)','User 2 (near user)');

figure;
plot(Pt, R1_av, 'linewidth', 1.5); hold on; grid on;
plot(Pt, R12_av, 'linewidth', 1.5);
plot(Pt, R2_av, 'linewidth', 1.5);
xlabel('Transmit power (dBm)');
ylabel('Achievable capacity (bps/Hz)');
legend('R_1','R_{12}','R_2')

Mô phỏng BER của NOMA trong kênh Rayleigh


clc; clear variables; close all;

N = 10^6;

d1 = 1000; d2 = 500; %Distances of users from base


station (BS)
a1 = 0.75; a2 = 0.25; %Power allocation factors
eta = 4; %Path loss exponent

%Generate rayleigh fading coefficient for both users


h1 = sqrt(d1^-eta)*(randn(1,N)+1i*randn(1,N))/sqrt(2);
h2 = sqrt(d2^-eta)*(randn(1,N)+1i*randn(1,N))/sqrt(2);

g1 = (abs(h1)).^2;
g2 = (abs(h2)).^2;
Pt = 0:2:40; %Transmit power in dBm
pt = (10^-3)*10.^(Pt/10); %Transmit power in linear scale
BW = 10^6; %System bandwidth
No = -174 + 10*log10(BW); %Noise power (dBm)
no = (10^-3)*10.^(No/10); %Noise power (linear scale)

%Generate noise samples for both users


w1 = sqrt(no)*(randn(1,N)+1i*randn(1,N))/sqrt(2);
w2 = sqrt(no)*(randn(1,N)+1i*randn(1,N))/sqrt(2);

%Generate random binary data for two users


data1 = randi([0 1],1,N); %Data bits of user 1
data2 = randi([0 1],1,N); %Data bits of user 2

%Do BPSK modulation of data


x1 = 2*data1 - 1;
x2 = 2*data2 - 1;

p = length(Pt);
for u = 1:p
%Do superposition coding
x = sqrt(pt(u))*(sqrt(a1)*x1 + sqrt(a2)*x2);
%Received signals
y1 = h1.*x + w1; %So phuc h1 * x + w1
y2 = h2.*x + w2;

%Equalize
eq1 = y1./h1; %So phuc y1 * x + w1
eq2 = y2./h2;

%AT USER 1--------------------


%Direct decoding of x1 from y1
x1_hat = zeros(1,N);
x1_hat(eq1>0) = 1;

%Compare decoded x1_hat with data1 to estimate BER


ber1(u) = biterr(data1,x1_hat)/N;

%----------------------------------

%AT USER 2-------------------------


%Direct decoding of x1 from y2
x12_hat = ones(1,N);
x12_hat(eq2<0) = -1;
y2_dash = eq2 - sqrt(a1*pt(u))*x12_hat;
x2_hat = zeros(1,N);
x2_hat(real(y2_dash)>0) = 1;

ber2(u) = biterr(x2_hat, data2)/N;


%-----------------------------------

gam_a =
2*((sqrt(a1*pt(u))-sqrt(a2*pt(u)))^2)*mean(g1)/no;
gam_b =
2*((sqrt(a1*pt(u))+sqrt(a2*pt(u)))^2)*mean(g1)/no;
ber_th1(u) = 0.25*(2 - sqrt(gam_a/(2+gam_a)) -
sqrt(gam_b/(2+gam_b)));

gam_c = 2*a2*pt(u)*mean(g2)/no;
gam_d = 2*((sqrt(a2) + sqrt(a1))^2)*pt(u)*mean(g2)/no;
gam_e = 2*((sqrt(a2) + 2*sqrt(a1))^2)*pt(u)*mean(g2)/no;
gam_f = 2*((-sqrt(a2) + sqrt(a1))^2)*pt(u)*mean(g2)/no;
gam_g = 2*((-sqrt(a2) +
2*sqrt(a1))^2)*pt(u)*mean(g2)/no;

gc = (1 - sqrt(gam_c/(2+gam_c)));
gd = (1-sqrt(gam_d/(2+gam_d)));
ge = (1-sqrt(gam_e/(2+gam_e)));
gf = (1-sqrt(gam_f/(2+gam_f)));
gg = (1-sqrt(gam_g/(2+gam_g)));

ber_th2(u) = 0.5*gc - 0.25*gd + 0.25*(ge+gf-gg);

gamma1(u) = a1*pt(u)*mean(g1)/(a2*pt(u)*mean(g1) + no);


%gamma = SNR
gamma2(u) = a2*pt(u)*mean(g2)/no;
end

semilogy(Pt, ber1,'r', 'linewidth',1.5); hold on; grid on;


semilogy(Pt, ber2,'b', 'linewidth',1.5);
semilogy(Pt, ber_th1, '*r','linewidth',1.5);
semilogy(Pt, ber_th2, '*b','linewidth',1.5);
xlabel('Transmit power (P in dBm)');
ylabel('BER');
legend('Sim. User 1/Far user','Sim. User 2/Near user','Theo.
User 1/Far user','Theo. User 2/Near user');
BER của NOMA khi phân bổ công suất cố định
clc; clear variables;

N = 5*10^5;
Pt = [40 20];
No = -114;

pt = (10^-3)*db2pow(Pt);
no = (10^-3)*db2pow(No);

a1 = 0:0.01:1;
a2 = 1 - a1;

d1 = 1000; d2 = 500;

eta = 4;
h1 = sqrt(d1^-eta)*(randn(1,N) + 1i*randn(1,N))/sqrt(2);
h2 = sqrt(d2^-eta)*(randn(1,N) + 1i*randn(1,N))/sqrt(2);

g1 = (abs(h1)).^2;
g2 = (abs(h2)).^2;

%Generate noise samples for both users


w1 = sqrt(no)*(randn(1,N)+1i*randn(1,N))/sqrt(2);
w2 = sqrt(no)*(randn(1,N)+1i*randn(1,N))/sqrt(2);

data1 = randi([0 1],1,N);


data2 = randi([0 1],1,N);

x1 = 2*data1 - 1;
x2 = 2*data2 - 1;
for v=1:2
for u = 1:length(a1)
x = sqrt(pt(v))*(sqrt(a1(u))*x1 + sqrt(a2(u))*x2);
y1 = x.*h1 + w1;
y2 = x.*h2 + w2;

%Equalize
eq1 = y1./h1;
eq2 = y2./h2;

%AT USER 1--------------------


%Direct decoding of x1 from y1
x1_hat = zeros(1,N);
x1_hat(eq1>0) = 1;

%Compare decoded x1_hat with data1 to estimate BER


ber1(u) = biterr(data1,x1_hat)/N;

%----------------------------------

%AT USER 2-------------------------


%Direct decoding of x1 from y2
x12_hat = ones(1,N);
x12_hat(eq2<0) = -1;

y2_dash = eq2 - sqrt(a1(u)*pt(v))*x12_hat;


x2_hat = zeros(1,N);
x2_hat(real(y2_dash)>0) = 1;

ber2(u) = biterr(x2_hat, data2)/N;


%-----------------------------------
end
if v==1
semilogy(a1, ber1,'-r', 'linewidth',1.5); hold on; grid
on;
semilogy(a1, ber2,'-b', 'linewidth',1.5);
xlabel('\alpha_1');
ylabel('BER');

end
if v==2
semilogy(a1, ber1,'--r', 'linewidth',1.5); hold on; grid
on;
semilogy(a1, ber2,'--b', 'linewidth',1.5);
xlabel('\alpha_1');
ylabel('BER');

end

end
legend('Far user p = 40','Near user p = 40','Far user p =
20','Near user p = 20');
So sánh phân bổ công suất cố định và công bằng
clc; clear variables; close all;

N = 10^5;
Pt = 30; %Max BS Tx power (dBm)
pt = (10^-3)*db2pow(Pt); %Max BS Tx power (Linear
scale)
No = -114; %Noise power (dBm)
no = (10^-3)*db2pow(No); %Noise power (Linear scale)

r = 0.5:0.5:10; %Far user target rate range


(R*)

df = 1000; dn = 500; %Distances

eta = 4; %Path loss exponent

p1 = zeros(1,length(r));
p2 = zeros(1,length(r));
pa1 = zeros(1,length(r));
pa2 = zeros(1,length(r));

af = 0.75; an = 0.25; %Fixed PA (for comparison)

hf = sqrt(df^-eta)*(randn(1,N) + 1i*randn(1,N))/sqrt(2);
hn = sqrt(dn^-eta)*(randn(1,N) + 1i*randn(1,N))/sqrt(2);

g1 = (abs(hf)).^2;
g2 = (abs(hn)).^2;

for u = 1:length(r)
epsilon = (2^(r(u)))-1; %Target SINR for far
user

%BASIC FAIR PA%


aaf = min(1,epsilon*(no +
pt*g1)./(pt*g1*(1+epsilon)));
aan = 1 - aaf;

% %IMPROVED FAIR PA%


% aaf = epsilon*(no + pt*g1)./(pt*g1*(1+epsilon));
% aaf(aaf>1) = 0;
% aan = 1 - aaf;

gamma_f = pt*af*g1./(pt*g1*an + no);


gamma_nf = pt*af*g2./(pt*g2*an + no);
gamma_n = pt*g2*an/no;

gamm_f = pt*aaf.*g1./(pt*g1.*aan + no);


gamm_nf = pt*aaf.*g2./(pt*g2.*aan + no);
gamm_n = pt*g2.*aan/no;

Cf = log2(1 + gamma_f);
Cnf = log2(1 + gamma_nf);
Cn = log2(1 + gamma_n);

Ca_f = log2(1 + gamm_f);


Ca_nf = log2(1 + gamm_nf);
Ca_n = log2(1 + gamm_n);

for k = 1:N
if Cf(k) < r(u)
p1(u) = p1(u) + 1;
end
if (Cnf(k)<r(u))||(Cn(k) < r(u))
p2(u) = p2(u) + 1;
end

if Ca_f(k) < r(u)


pa1(u) = pa1(u) + 1;
end
if aaf(k) ~= 0
if (Ca_n(k) < r(u)) || (Ca_nf(k) < r(u))
pa2(u) = pa2(u) + 1;
end
else
if Ca_n(k) < r(u)
pa2(u) = pa2(u) + 1;
end
end
end
end
pout1 = p1/N;
pout2 = p2/N;
pouta1 = pa1/N;
pouta2 = pa2/N;

figure;
plot(r,pout1,'--+r','linewidth',2); hold on; grid on;
plot(r,pout2,'--ob','linewidth',2);
plot(r,pouta1,'r','linewidth',2);
plot(r,pouta2,'b','linewidth',2);
xlabel('Target rate of far user (R*) bps/Hz');
ylabel('Outage probability');
xlim([r(1) r(end)]);
legend('Far user (fixed PA)','Near user (fixed PA)','Far
user (fair PA)','Near user (fair PA)');

Mã hóa chồng chất

clc; clear all; close all;

x1 = [1 0 1 0];
x2 = [0 1 1 0];

xmod1 = 2*x1-1;
xmod2 = 2*x2-1;

a1 = 0.75; a2 = 0.25;
x = sqrt(a1)*xmod1 + sqrt(a2)*xmod2;

%Plot figures

ay = -2:0.2:2;
ax = ones(1,length(ay));

figure;
subplot(2,1,1)
stairs([x1,x1(end)],'linewidth',2);
ylim([-2 2])
grid on; hold on;
title('Data of user 1 (x_1)')
for u = 1:3
plot(ax*(u+1),ay,':k','linewidth',2);
end
subplot(2,1,2)
stairs([x2,x2(end)],'m','linewidth',2);
ylim([-2 2])
grid on; hold on;
for u = 1:3
plot(ax*(u+1),ay,':k','linewidth',2);
end
title('Data of user 2 (x_2)')

figure;
subplot(2,1,1)
stairs([xmod1,xmod1(end)],'linewidth',2);
ylim([-2 2])
grid on; hold on;
title('Data of user 1 (x_1)')
for u = 1:3
plot(ax*(u+1),ay,':k','linewidth',2);
end
subplot(2,1,2)
stairs([xmod2,xmod2(end)],'m','linewidth',2);
ylim([-2 2])
grid on; hold on;
for u = 1:3
plot(ax*(u+1),ay,':k','linewidth',2);
end
title('Data of user 2 (x_2)');

t1 = sqrt(a1)*xmod1;
t2 = sqrt(a2)*xmod2;
figure;
subplot(2,1,1)
stairs([t1,t1(end)],'linewidth',2);
ylim([-2 2])
grid on; hold on;
title('Scaled data of user 1 ($$\sqrt{a_1}x_1$
$)','Interpreter','latex','FontSize',13)
for u = 1:3
plot(ax*(u+1),ay,':k','linewidth',2);
end
subplot(2,1,2)
stairs([t2,t2(end)],'m','linewidth',2);
ylim([-2 2])
title('Scaled data of user 2 ($$\sqrt{a_2}x_2$
$)','Interpreter','latex','FontSize',13)
grid on; hold on;

for u = 1:3
plot(ax*(u+1),ay,':k','linewidth',2);
end

figure;
stairs([x,x(end)],'r','linewidth',2);
grid on; hold on;
for u = 1:3
plot(ax*(u+1),ay,':k','linewidth',2);
end
title('Superposition coded signal')

Loại bỏ nhiễu liên tiếp (SIC)

clc; clear all; close all;

x1 = [1 0 1 0];
x2 = [0 1 1 0];

xmod1 = 2*x1-1;
xmod2 = 2*x2-1;

a1 = 0.75; a2 = 0.25;
x = sqrt(a1)*xmod1 + sqrt(a2)*xmod2;

xdec1 = ones(1,length(x1));
xdec1(x<0)=-1;

xrem = x - sqrt(a1)*xdec1;
xdec2 = zeros(1,length(x1));
xdec1(x<0)=0;
xdec2(xrem>0)=1;

%Plot figures

ay = -2:0.2:2;
ax = ones(1,length(ay));

figure;
subplot(2,1,1)
stairs([x1,x1(end)],'linewidth',2);
ylim([-2 2])
grid on; hold on;
title('Data of user 1 (x_1)')
for u = 1:3
plot(ax*(u+1),ay,':k','linewidth',2);
end
subplot(2,1,2)
stairs([x2,x2(end)],'m','linewidth',2);
ylim([-2 2])
grid on; hold on;
for u = 1:3
plot(ax*(u+1),ay,':k','linewidth',2);
end
title('Data of user 2 (x_2)')

figure;
subplot(2,1,1)
stairs([xmod1,xmod1(end)],'linewidth',2);
ylim([-2 2])
grid on; hold on;
title('Data of user 1 (x_1)')
for u = 1:3
plot(ax*(u+1),ay,':k','linewidth',2);
end
subplot(2,1,2)
stairs([xmod2,xmod2(end)],'m','linewidth',2);
ylim([-2 2])
grid on; hold on;
for u = 1:3
plot(ax*(u+1),ay,':k','linewidth',2);
end
title('Data of user 2 (x_2)');

t1 = sqrt(a1)*xmod1;
t2 = sqrt(a2)*xmod2;
figure;
subplot(2,1,1)
stairs([t1,t1(end)],'linewidth',2);
ylim([-2 2])
grid on; hold on;
title('Scaled data of user 1 ($$\sqrt{a_1}x_1$
$)','Interpreter','latex','FontSize',13)
for u = 1:3
plot(ax*(u+1),ay,':k','linewidth',2);
end
subplot(2,1,2)
stairs([t2,t2(end)],'m','linewidth',2);
ylim([-2 2])
title('Scaled data of user 2 ($$\sqrt{a_2}x_2$
$)','Interpreter','latex','FontSize',13)
grid on; hold on;

for u = 1:3
plot(ax*(u+1),ay,':k','linewidth',2);
end

figure;
stairs([x,x(end)],'r','linewidth',2);
grid on; hold on;
for u = 1:3
plot(ax*(u+1),ay,':k','linewidth',2);
end
title('Superposition coded signal')
plot(1:5,zeros(1,5),'k','linewidth',1.5)

figure;
stairs([xrem,xrem(end)],'r','linewidth',2);
grid on; hold on;
for u = 1:3
plot(ax*(u+1),ay,':k','linewidth',2);
end
title('Superposition coded signal')
plot(1:5,zeros(1,5),'k','linewidth',1.5)

Common questions

Powered by AI

The path loss exponent plays a critical role in determining the performance of NOMA systems by affecting both channel capacity and BER. A higher path loss exponent implies greater signal degradation over distance, reducing channel capacity and increasing BER, especially for users located further from the transmission source. This necessitates adjustments in power allocation strategies to mitigate such performance drops, thus emphasizing the importance of adaptive power control in NOMA systems .

In NOMA systems, the fixed power allocation strategy assigns predetermined power levels to users regardless of their channel conditions, typically offering more power to weaker channels to satisfy the fairness constraint. Conversely, the fair power allocation dynamically adjusts power levels based on the users' channel conditions, optimizing performance by possibly allowing greater throughput for channels in favorable conditions. Fair allocation may lead to reduced outage probabilities since it allocates resources more adaptively compared to the static distribution of fixed power allocation .

The capacity in NOMA systems improves with increased transmit power due to the power-domain multiplexing, which enables simultaneous transmission to multiple users. This leads to higher data rates as the signal strength increases. In OMA systems, capacity also increases with transmit power but is inherently limited by the fact that users share the channel in different time slots, potentially reducing the total capacity compared to NOMA. Simulations show that NOMA typically offers a higher capacity than OMA across varying transmit power levels .

BPSK modulation impacts NOMA performance by providing a straightforward, robust method for digital signal modulation that supports low BER in high SNR scenarios. In NOMA, where multiple users’ data is superimposed for simultaneous transmission, BPSK aids in more reliable signal demodulation at the receiver end due to its simplicity and effectiveness in mitigating errors over noisy channels, such as those influenced by AWGN .

User distance from the base station significantly impacts the achievable capacity in NOMA systems. As distance increases, signal attenuation due to path loss becomes more pronounced, leading to reduced channel capacity for far users. To counteract this, NOMA typically allocates higher power to far users compared to near users to equalize capacity among users. Thus, distance-related degradation necessitates strategic power allocation to maintain system fairness and capacity .

Successive Interference Cancellation (SIC) is crucial in NOMA systems as it allows for the separation and decoding of superimposed signals from different users. This technique leverages differences in power levels to decode and remove stronger signals, thereby minimizing interference for weaker signals. Effectively implemented SIC improves the performance of NOMA systems by reducing error rates and enhancing throughput, but its complexity can increase decoding delays and computational load .

In an AWGN channel, the BER varies between near and far users in NOMA due to differences in received power levels. Near users typically enjoy stronger signals due to lesser path loss, resulting in lower BER. Far users, however, receive significantly attenuated signals and experience higher BER, necessitating greater allocated power to maintain acceptable performance. This disparity is due to the distance-dependent attenuation effects inherent in wireless transmission .

Rayleigh fading significantly impacts the Bit Error Rate (BER) in NOMA systems by introducing variability and signal degradation, particularly for users with weaker channels or those farther from the base station. This effect increases the BER, as random fluctuations due to fading create discrepancies between transmitted and received signals. Users positioned farther experience higher BER readings because of additional signal attenuation caused by increased path loss and fading effects .

SC-NOMA demonstrates a higher sum rate performance compared to TDMA under varying SNR conditions. The sum rate for SC-NOMA is calculated as a function of the power allocation coefficients and SNR, resulting in generally higher achievable rates than TDMA, where users share time and thus, the channel, leading to divided capacity. The simulation results clearly show that SC-NOMA outperforms TDMA across different SNR levels .

Fixed power allocation involves lesser computational complexity as it requires setting static power levels, irrespective of channel conditions, simplifying the system. However, this simplicity can result in higher outage probabilities, especially in varying channel conditions. In contrast, fair power allocation, while more computationally intensive due to its adaptive nature, optimizes power distribution dynamically based on channel conditions, thereby reducing outage probabilities and enhancing system resilience against channel variability .

You might also like