A New Methodology for Radio Spectrum Requirement Estimation of Wireless Communication Systems

This thesis presents a new methodology for estimating the required amount of radio spectrum for wireless communication systems. The estimation is
based on assumptions on the offered traffic and traffic characteristics as well as the capabilities and deployment of the available radio access networks.
The methodology was originally developed in the scope of the preparations towards the ITU-R World Radiocommunication Conference 2007. After the
finalization of the ITU standardization the author developed the methodology further. The main improvements were made in the scenario definition
framework, the traffic distribution algorithm and the capacity dimensioning procedure for packet-switched services. By making a clear separation
between methodology specification and its applications in terms of scenario development and the actual calculation the specific ITU methodology was
converted into a generic tool that can be applied to any scenario. In addition, modifications within the calculation of spectrum requirements from capacity
requirements were made (e.g. differentiation between FDD and TDD based systems). The methodology’s packet-switched capacity calculation algorithm involves two different approximations for the IP packet waiting time and delay distribution. One is based on a hyper-exponential distribution and the other
one is based on the Gamma distribution. A detailed performance evaluation of these approximations shows that both approximation methods are very
accurate. Their main difference is in the computational effort required and their applicability throughout the whole parameter space. The methodology comprises two algorithms for calculating the system capacity required to serve the given offered traffic under given Quality of Service (QoS) constraints. While the approach and the algorithm for circuitswitched traffic are based on the theory of multi-service loss models, which has been extensively covered in previous work, concept and algorithm for packet-switched traffic are completely novel. Therefore, an extensive and systematic sensitivity analysis of the packet-switched capacity calculation algorithm is presented in order to illustrate the behavior of the model.