Accuracy of local mean signal prediction in urban environments: input database granularity issues

A number of propagation prediction tools are being developed to support urban mobile radio planning. A significant number of these are based on ray-optical techniques. Sources of prediction error in such propagation models can be broadly categorised into two areas: (i) electromagnetic modelling errors and (ii) errors introduced through the imprecise description of the input environment. This paper proposes to examine the dependence of the local mean signal scattered from building walls and corners on the level of spatial detail input into a three-dimensional ray-tracing propagation tool.