Spectrum access in white spaces using spectrum sensing and geolocation databases

A spectrum license grants users the right to transmit on a particular piece of spectrum. Historically, a spectrum license has been allocated for a particular technology. While this strategy facilitates interference control, it also results in spectrum scarcity as more spectrum-efficient technologies are invented. In order to meet the increasing data traffic demands in a timely manner, a shared use of the spectrum seems to be the only viable solution. According to this line of thinking, different technologies with possibly different deployment densities can share the same spectrum under certain conditions. While shared spectrum access improves spectral efficiency, it also increases the risk for harmful interference among the different systems. This calls for a change in the traditional way of issuing spectrum licenses: instead of specifying transmit power levels, the spectrum usage rights specify the generated interference that is permitted. Spectrum access to white spaces would enhance spectrum utilisation, while also testing the approach of controlling the interference between different systems directly rather than through the transmission power. The amount of interference generated to the license holder can be controlled by spectrum sensing and/or geolocation database access. Interference control using spectrum sensing usually boils down to a signal detection problem. In this thesis, we show that the traditional signal detection framework is not appropriate for recovering transmission opportunities in the spatial domain and propose an alternative model. Also, sensing strategies for energy efficient wideband spectrum sensing and trade-off analysis between the service requirement and the demand in the measured spectrum are demonstrated. At this moment, spectrum access to white spaces is mostly possible via geolocation databases. The database is responsible for handling spectrum access requests while complying with certain regulatory conditions. In this thesis, we suggest some interference control and power allocation algorithms that may govern the operation of the database. The algorithms have a low complexity to enable a real-time operation in the database. They involve simple models to capture the impact of the non-uniform demand density, terrain-based propagation and fading correlations on the generated interference. Also, we propose a joint rate and power allocation algorithm that protects the license holder in all cases.