On the usage of geolocation aware spectrum measurements in spectrum sharing

Over the last few decades wireless technology has come a long way considering the enhancement in throughput and quality of service. Although the engineered solutions tend to have well defined bounds, the wireless users' traffic is limited only by their imagination. Spectrum sharing has been propsed as a key enabling technology to meet the demand of future wireless systems. Licensed Shared Access and Spectrum Access System, spectrum sharing standardisation efforts, have identified the importance of determining the incumbent exclusion zone in geographical space. In this regard this thesis proposes to use geolocation aware spectrum measurements to estimate the incumbent exclusion zone by determining the incumbent location and the transmit power. The thesis formulates the task in hand as an underdetermined ℓ₁-norm minimisation problem that leverages the incumbent characteristics. With a commercial intent we keep the complexity to a minimum by considering a discrete space model. We analytically prove that the channel state matrix in this problem meets the Restricted Isometry Property, a requirement to obtain a unique solution for an underdetermined problem. We develop algorithms that can effectively determine the incumbent exclusion zones in different fading environments. We further consider the scenario of having off-grid nodes in or discrete space model. We capture this effect as a mismatch in the channel state matrix. We formulate a linear algebra based solution to overcome this error without increasing the complexity of the solution and the traffic overhead. We analytically find the necessary condition that should be satisfied by any approach attempting to eliminate this error, and find the upper bound of the probability of this condition being satisfied. We apply this framework to the Licensed Share Access cellular network architecture. We develop a downlink power allocation algorithm that carefully controls the power based on the exclusion zones determined using this framework. We further strengthen the coupling of our framework with Licensed Share Access by developing a Kalman Filter based approach to determine the exclusion zone based on incumbent features provided by the Licensed Share Access repository. In summary, geolocation awareness in wireless networks has not been well studied. This thesis has given rise to a framework based on geolocation aware spectrum measurements that can be used in the context of spectrum sharing. It is believed that this framework will make a contribution to spectrum sharing standardisation efforts in the near future.