how can eliminate negative data in erf function?

Question

leuva 2023-3-10

0
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/1926250-how-can-eliminate-negative-data-in-erf-function

评论： leuva 2023-3-14

采纳的回答： Alan Stevens

clc;

clear all;

close all;

N = 30 ;

mu =N*pi/4;

M = N*(1-(pi/16));

lemdaD = 2;

avgsnr = 0.0001;

p = 100

n = 3.5;

deld =1e-3 * 10^(0/10); % 0 dBm in watts;

delr = 10^(-80/10)/1000; % -80 dBm in watts

delrd =deld+(p^2)*delr;

w = [0:1:40]

ps = zeros(size(w));

for i = 1:length(w)

%for j= 1:1:5

ps(i) = 10.^(w(i)./10)./1000;

dsr = 41.23;

drd = 60.82;

dsd =100;

d0 =1

avgsnrd(i) = (ps(i)./delrd);

b1(i) =sqrt(avgsnrd(i)).*p*(dsr*drd/d0^2)^(-0.5*n)

b2(i) = sqrt(avgsnrd(i)).*(dsd/d0)^(-0.5*n)

v1(i) = 1./b2(i)

v2(i) = b1(i)./b2(i)

e1(i) = sqrt(M.*v2(i)+lemdaD)

e2(i) = sqrt(2*M*lemdaD)

e3(i) = M.*v1(i).*v2(i)

e4(i) = (lemdaD.*v1(i))./v2(i)

e5 = lemdaD*mu

sqrt_avgsnr = sqrt(avgsnr)

d1(i)= 1./b1(i)

term1(i) = 0.5 .* (erf((d1(i).*sqrt_avgsnr - mu)./ sqrt(2* M)) + ...

erf(mu/sqrt(2 * M)))

term2(i) = (sqrt(lemdaD)./(2.*e1(i))).*exp(-(v1(i).*sqrt_avgsnr - v2(i).* mu).^2./(2.*e1(i).^2)).*....

erf((e4(i).*sqrt_avgsnr - e5)./(e1(i) .* e2(i)))

term3(i)= (sqrt(lemdaD)./(2.*e1(i))).*exp(-(v1(i).*sqrt_avgsnr - v2(i).* mu).^2./(2.*e1(i).^2)).* ....

erf((e3(i).*sqrt_avgsnr + e5)./(e1(i) .* e2(i)))

FyD1(i) = term1(i) - term2(i) - term3(i)

%FyD11(i) = abs(FyD1(i))

end

semilogy(w, max(FyD1, 0), '-or')

legend('p = 0 dB')

%,'p = 10 dB','p = 15 dB','p = 20 dB')

xlabel('ps(dBm)')

ylabel('Pout')

2 个评论
显示无隐藏无

Arka 2023-3-10

编辑：Arka 2023-3-10

Do you want to remove the negative data points from the term1, term2, and term3 vectors?

leuva 2023-3-10

yes

请先登录，再进行评论。

请先登录，再回答此问题。

Answer 1

Alan Stevens 2023-3-10

0
链接

此回答的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/1926250-how-can-eliminate-negative-data-in-erf-function#answer_1190380

在 MATLAB Online 中打开

Should

term2(i) = (sqrt(lemdaD)./(2.*e1(i))).*exp(-(v1(i).*sqrt_avgsnr - v2(i).* mu).^2./(2.*e1(i).^2)).*....
        erf((e4(i).*sqrt_avgsnr - e5)./(e1(i) .* e2(i)));

be

term2(i) = (sqrt(lemdaD)./(2.*e1(i))).*exp(-(v1(i).*sqrt_avgsnr - v2(i).* mu).^2./(2.*e1(i).^2)).*....
        erf((e4(i).*sqrt_avgsnr + e5)./(e1(i) .* e2(i)));

i.e. have a " + e5" instead of " - e5"?

This removes the negatives.