0
$\begingroup$

I was reading a paper about communications between nodes, and in the simulation section they gave the condition of the simulation as follow :

"...The noise power is assumed to be the same at all nodes..."

In the simulation it is never stated that the transmit power of all nodes are not the same, but if you put them so that they are equal you get an error. My guess is that they mean the noise variances $σ^2$ of the nodes are the same. However the simulation result is way off the mark from the result in the paper.

Did I understand their meaning properly?

Improve this question
$\endgroup$
7
  • $\begingroup$ Yes, it means the noise variance is the same. Can you give more details about the simulations in the paper and how you approached it? $\endgroup$
    – BlackMath
    Commented Apr 6, 2019 at 16:51
  • $\begingroup$ @BlackMath The simulations is from "DEVICE-TO-DEVICE COMMUNICATIONS: THE PHYSICAL LAYER SECURITY ADVANTAGE" about measuring secrecy outage probability (SOP) between two modes, one is through a relay R and the other directly between two users A and B, with the present of eavesdropper E. $\endgroup$
    – Kodiak
    Commented Apr 7, 2019 at 1:21
  • $\begingroup$ The paper derived two formula to calculate the SOPs, based on the location of the eavesdropper. The model based on the fact that SNR is propotional to distance power path loss exponent, and disproportional to noise variance.They assumed that the distances between A, B, and R, and between E and R is known; all SNR between A, B, R is equal, known path loss exponent, target secret rate = 1, number of antennas from R and E is known. Then they find the SOPs as a function of SNR and the SOPs as a function of number of antenna R have. $\endgroup$
    – Kodiak
    Commented Apr 7, 2019 at 1:23
  • $\begingroup$ I followed the mathematical formula, assuming the transmit power P is given. With known path loss, known distance, I calculate the noise variance value based on given average SNR pAB , which is shared between all nodes, thus find out the according average SNR pAE and pRE. This allow me to map out a very complicated formula f(pAE,pRE,pAB,NR) (NR is the number of antenna R have). In the paper they always have unity SOP values except when E is in an extreme location. In my simulation they have a weird curve, that SOP is prone to attack when SNR level close to 0 but drop in every other SNR. $\endgroup$
    – Kodiak
    Commented Apr 7, 2019 at 1:36
  • $\begingroup$ I was thinking of sharing the formulas, and how you simulated them in a code, instead of explaining the system model. $\endgroup$
    – BlackMath
    Commented Apr 7, 2019 at 12:23

1 Answer 1

Reset to default
1
$\begingroup$

Although noise is often called channel noise in various block diagrams of radio-frequency (RF) communications systems that communications analysts draw, the truth of the matter is that most actual RF channels are mostly noiseless, and what is referred to as the channel noise is primarily thermal noise in the resistive elements in the front end of the receiver. The noise power in the receiver that communications engineers have to contend with is thus determined by the temperature of those resistors (and the value of those resistors). Thus, assuming that the receivers in the various "nodes" are identical and also in the same geographical area so that all the receivers can be expected to be at the same temperature, it is reasonable to use a model in which the noise power is the same in all of the receivers.

Improve this answer
$\endgroup$
3
  • $\begingroup$ The strange thing is that throughout the paper and the analytical formula they never mentioned noise power at all. The only thing about noise they included is noise variance. In the model building they stated that the two nodes they want to observe had "spatial white noise with variance x and y". $\endgroup$
    – Kodiak
    Commented Apr 7, 2019 at 1:37
  • $\begingroup$ @Kodiak They obviously mean that they have already translated continuous-time RF signals into discrete-time baseband equivalents and so the noise power means the same as the noise variance. $\endgroup$
    – Dilip Sarwate
    Commented Apr 7, 2019 at 2:26
  • $\begingroup$ Thank you for the explanation! $\endgroup$
    – Kodiak
    Commented Apr 7, 2019 at 8:17

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.