| Author | Pati, Bipun Man |
| Call Number | AIT Thesais no.TC-16-03 |
| Note | A thesis submitted in partial ful llment of the requirements for the
degree of Master of Engineering in
Telecommunications, School of Engineering and Technology |
| Publisher | Asian Institute of Technology |
| Abstract | Spectrum sensing is one of the most crucial task of Cognitive radio (CR). The
cognitive radio consists of Low noise ampli er (LNA). For devices like cognitive ra-
dio which requires low cost mass production and high power eciency, it is evident
that operation of LNA becomes highly non-linear. Due to non-linearity of LNA we
get Intermodulation (IM) product at the output of LNA. The e ect of second order
non-linearity can be ltered out while the e ect of third order non-linearity cannot be
ltered out in wide band spectrum sensing. We show that the degradation due to third
order non-linearity has severe impact on the performance of cognitive radio resulting
in low detection probability. We show that the third order non-linearity generated by
the LNA seems to t the Cauchy distribution, but for simplicity and mathematical
analysis we model the e ect of third order non-linearity as Gaussian distribution. We
then perform the theoretical analysis based on this approximation and derive the ex-
pression for detection and false alarm probability. We then analyze the performance
of energy detector in a CR due to the interference from other cognitive radios. We
model the generation of CR and their interference using Poisson eld of interferers.
For simplicity, we use energy detector for spectrum sensing. Energy detector deter-
mines the occupancy of a channel by comparing the energy in the particular band with
the threshold value which is predetermined. The expression for the threshold selec-
tion are also analyzed based on Gaussian approximation. The threshold selection is
done based on the assumption of noise as Gaussian process which seems reasonable
when we do not consider the e ect of non-linearity. However, when the e ect of LNA
non-linearity is taken into account we show that the threshold selection is actually
summation of correlated Gaussian random variable with Gaussian random variable.
We approximate this summation as multivariate Gaussian mixture model. Finally, we
perform simulation. At rst, we show simulation result investigating the performance
of energy detector with only the e ect of non-linearity and then, we show the result
for the case of Poisson eld of interferers. The thesis concludes by showing how the
work in this thesis can be extended for further research. |
| Year | 2016 |
| Type | Thesis |
| School | School of Engineering and Technology (SET) |
| Department | Department of Civil and Infrastucture Engineering (DCIE) |
| Academic Program/FoS | Telecommunications (TC) |
| Chairperson(s) | Attaphongse Taparugssanagorn; |
| Examination Committee(s) | Teerapat SanguankotchakornShinichi Nakamura; |
| Scholarship Donor(s) | Asian Development Bank-Japan Scholarship Program (ADB-JSP); |