Abstract:
The rapid advancement in the wireless technology has created scarcity of the spectrum
bandwidth. The wireless connectivity is rapidly increasing which also involves the
crowding of unlicensed spectra. This scarcity of spectra has pushed the regulatory
authorities to device new ways for the efficient utilization of the existing spectrum.
According to Federal Communication Commission (FCC) more than 70% of the
available spectrum is not utilized optimally. Due to the shortage of available frequencies
the bandwidth becomes expensive. For optimal and efficient usage of spectrum one
possibility is to scan the whole spectrum to determine the opportunity for transmission.
The term Cognitive Radio refers to the intelligent radios that have spectrum scanning and
parameter adjustment capability. This thesis presents schemes for intra-network
spectrum sharing in centralized cognitive radio networks. In such schemes a centralized
spectrum server is responsible for sharing the spectrum among the cognitive radio users.
All transmitters are assumed to have fixed transmitting power. The data rate for the links
is computed by using the signal-to-interference ratio on those links. The centralized
spectrum server is assumed to have the prior knowledge about the link gains based on
which it finds an optimal schedule that maximizes the average data rate on all the active
links. Three important scheduling schemes Maximum Sum Rate Scheduling, Max-min
Fair Scheduling and Proportional Fair scheduling have been implemented in the context
of cognitive radio networks. All of these techniques try to maximize different parameters
with the sole objective of maximizing the utility of spectra. In the end Dynamic
Scheduling is done by using the three above mentioned techniques by taking the
maximum data need as input from the CR users. A simulation of these techniques has
been developed in MATLAB and a thorough comparison of these techniques has also
been performed.
