In the field of spectrum sensing for cognitive radios, the cyclostationary features detector is recommended to improve the performance in terms of probability of detection, compared to the energy detector and others. However, I have not found analytical expressions (i.e., Pd vs. SNR, Pd vs. Pfa) associated to this technique.
Actually I found one paper on spectrum sensing and they use one expression that depends of the noise power of the channel. However, for AWG noise, it is supposed that no cyclic features are present, so the performance of the cyclostationary features detector shouldn´t depend on noise power. Am I wright?
By computing the cyclic autocorrelation function (CAF), the noise will get value only at cyclic frequency= zero. Meanwhile, the signals, for example SU signal, is expected to get value higher than zero at cyclic frequency> zero. Furthermore, the wireless channel is not only AWGN it also face fading. Fading will effect on Pd results.
There are many researches that discussed Pd vs. SNR and Pd vs. Pfa for cyclostationary feature detection. I cited them in my research works, you can check my publications, such as ,
- A.V. Dandawate, G.B. Giannakis, "Statistical tests for presence of cyclostationarity," Signal Processing, IEEE Transactions on , vol.42,no.9, pp.2355-2369, Sep 1994.
By computing the cyclic autocorrelation function (CAF), the noise will get value only at cyclic frequency= zero. Meanwhile, the signals, for example SU signal, is expected to get value higher than zero at cyclic frequency> zero. Furthermore, the wireless channel is not only AWGN it also face fading. Fading will effect on Pd results.
There are many researches that discussed Pd vs. SNR and Pd vs. Pfa for cyclostationary feature detection. I cited them in my research works, you can check my publications, such as ,
- A.V. Dandawate, G.B. Giannakis, "Statistical tests for presence of cyclostationarity," Signal Processing, IEEE Transactions on , vol.42,no.9, pp.2355-2369, Sep 1994.
Mr. Alfaqawi,
Thanks for your answer. I do have that paper of Giannakis, but I tried once to implement their statistical tests and they became so computationally expensive that it was not useful for practical applications. Anyway, I will check it again.
I would like to know if you have used their results in your work. I will take a look to your papers.
Thank you again.
Hi,
In couple of month before I also had same question regarding analytical expression for cyclostationary detection. Almost all papers are in non parametric sense. So I satisfy myself that it is not possible to determine the analytical expression for Pd and Pf through normal way. Although through invoking some approximation it is possible to evaluate the expression but that expression cannot be used in generalized manner so it is worthless to use that.
Apart from that the paper you are talking about is most probably this one
"Cyclostationary Property Based Spectrum Sensing Algorithms for
Primary Detection in Cognitive Radio Systems".
You have taken a very good point and upto my knowledge you are correct. But still its expression is correct in limited sense if you consider your signal power as unity. So instead of considering only noise power you will consider SNR with same expression.
I hope done answer to your question...
- Badges
- Science topic