Generalised Radio Resource Sharing Framework for Heterogeneous Radio Networks

Recent years have seen a significant interest in quantitative measurements of licensed
and unlicensed spectrum use. Several research groups, companies and regulatory bodies
have conducted studies of varying times and locations with the aim to capture the over-
all utilisation rate of spectrum. The studies have shown that large amount of allocated
spectrum are under-utilised, and create the so called \spectrum holes", resulting in a
waste of valuable frequency resources. In order to satisfy the requirements of increased
demands of spectrum resources and to improve spectrum utilisation, dynamic spectrum
sharing (DSS) is proposed in the literature along with cognitive radio networks (CRNs).
DSS and CRNs have been studied from many perspectives, for example spectrum sensing
to identify the idle channels has been under the microscope to improve detection proba-
bility. As well as spectrum sensing, the DSS performance analysis remains an important
topic moving towards better spectrum utilisation to meet the exponential growth of
traffi�c demand. In this dissertation we have studied both techniques to achieve different
objectives such as enhancing the probability of detection and spectrum utilisation.
In order to improve spectrum sensing decisions we have proposed a cooperative spec-
trum sensing scheme which takes the propagation conditions into consideration. The
proposed location aware scheme shows an improved performance over conventional hard
combination scheme, highlighting the requirements of location awareness in cognitive
radio networks (CRNs).
Due to the exponentially growing wireless applications and services, traffi�c demand is
increasing rapidly. To cope with such growth wireless network operators seek radio
resource cooperation strategies for their users with the highest possible grade of service
(GoS). However, it is diffi�cult to fathom the potential benefits of such cooperation, thus
we propose a set of analytical models for DSS to analyse the blocking probability gain and
degradation for operators. The thesis focuses on examining the performance gains that
DSS can entail, in different scenarios. A number of dynamic spectrum sharing scenarios
are proposed. The proposed models focus on measuring the blocking probability of
secondary network operators as a trade-o� with a marginal increase of the blocking
probability of a primary network in return of monetary rewards. We derived the global
balance equation and an explicit expression of the blocking probability for each model.
The robustness of the proposed analytical models is evaluated under different scenarios
by considering varying tra�c intensities, different network sizes and adding reserved
resources (or pooled capacity). The results show that the blocking probabilities can
be reduced significantly with the proposed analytical DSS models in comparison to the
existing local spectrum access schemes.
In addition to the sharing models, we further assume that the secondary operator aims
to borrow spectrum bandwidths from primary operators when more spectrum resources
available for borrowing than the actual demand considering a merchant mode. Two
optimisation models are proposed using stochastic optimisation models in which the secondary operator (i) spends the minimum amount of money to achieve the target
GoS assuming an unrestricted budget or (ii) gains the maximum amount of pro�t to
achieve the target GoS assuming restricted budget. Results obtained from each model
are then compared with results derived from algorithms in which spectrum borrowings
were random. Comparisons showed that the gain in the results obtained from our pro-
posed stochastic optimisation model is significantly higher than heuristic counterparts.
A post-optimisation performance analysis of the operators in the form of analysis of
blocking probability in various scenarios is investigated to determine the probable per-
formance gain and degradation of the secondary and primary operators respectively.
We mathematically model the sharing agreement scenario and derive the closed form
solution of blocking probabilities for each operator. Results show how the secondary
and primary operators perform in terms of blocking probability under various offered
loads and sharing capacity.
The simulation results demonstrate that at most trading windows, the proposed opti-
mal algorithms outperforms their heuristic counterparts. When we consider 80 cells,
the proposed pro�t maximisation algorithm results in 33.3% gain in net pro�t to the
secondary operators as well as facilitating 2.35% more resources than the heuristic ap-
proach. In addition, the cost minimisation algorithm results in 46.34% gain over the
heuristic algorithm when considering the same number of cells (80).