I wanna add (2 dB for example) of noise uncertainty to a given signal x, I tried this way:
S=10*log10(var(x));
N=S-SNR;
n1=randn(1,round(length(x)/4));
n1=n1-mean(n1);
n1=n1/std(n1)*sqrt(10^(N/10));
n2=randn(1,round(length(x)/4));
n2=n2-mean(n2);
n2=n2/std(n2)*sqrt(10^((N+1)/10));
n3=randn(1,round(length(x)/4));
n3=n3-mean(n3);
n3=n3/std(n3)*sqrt(10^(N/10));
n4=randn(1,length(sf)-3*length(n1));
n4=n4-mean(n4);
n4=n4/std(n4)*sqrt(10^((N+2)/10));
n=[n1 n2 n3 n4];
I assumed that the vector "n" will add a 2 dB of uncertainty to the clean signal x to produce y as follow:
y=x+n;
Is that the right way to add uncertainty?
Best
If you are using matlab and you wish to add AWGN, have a look at this function:
http://au.mathworks.com/help/comm/ref/awgn.html
or follows this discussion:
http://stackoverflow.com/questions/23690766/proper-way-to-add-noise-to-signal
Hope this helps.
Thanks very much for sharing these links, essentially, what I'm looking for is generating a fluctuating Gaussian noise (non-stationary) that has varying noise power in the range of (1+p)-1 and (1+p) where p is the uncertainty factor in dB.
Thank you very much for reply, theoretically, I know the boundaries of uncertainty, but my problem still unsolved, because I want to program the uncertainty using MATLAB in order to test under some algorithms of spectrum sensing.
Dear Prof. Guimarães
Let me express my deep appreciation for the valuable feedback.
Best regards
https://www.researchgate.net/publication/258835771_Decision_making_under_uncertainty_in_energy_systems_State_of_the_art
Article Decision making under uncertainty in energy systems: State of the art
- Badges
- Science topic
- Similar topics
- Engineering
- Electrical Engineering
- Signal Process