Spatial statistics and models of spectrum use

In order to opportunistically exploit unused radio spectrum nodes of dynamic spectrum access (DSA) networks monitor the spectrum around them. Such cognitive radios can greatly benefit from a spatial characterization of spectrum use. However, there is need to find an efficient way to describe spatial use, something which has not been studied in details so far. In this paper, we introduce spatial statistics techniques as promising methods to describe spectrum use and enable optimization of DSA networks. We discuss two approaches to spatial modelling of spectrum, namely a deterministic approach based on a system model of the complete radio environment and an empirical approach that exploits passive measurements of the spectrum use. We elaborate on the impact of different network properties on the models and provide realistic parameter sets for generation of simulation scenarios. Additionally, we investigate cooperative sensing as a use case for spatial statistics based runtime optimization of the network configuration.