A GIS approach for mobile telephone signal path-loss prediction

Signal propagation path-loss prediction is a fundamental problem for the planning and design of a cellular mobile telephone system. Many models, empirical, experimental, or analytical, have been designed and used. The analytical models apply the path-loss prediction formulae to path profiles derived from a digital elevation model to obtain the signal path-loss. The digital elevation model stores detailed terrain height data. The heights can be sampled regularly, or selectively stored to capture important terrain features. The regular sampled elevation database (grid model) has been used for most of the existing analytical path loss prediction models. Carefully examination of the grid model reveals that it is not the only terrain representation. TIN (Triangular Irregular Network) terrain model can be used as an alternative elevation database. With the TIN model, sudden changes and other surface features that are difficult to store in the grid model can be included in the database. More accurate path profiles could be derived and more accurate path loss prediction would be achieved. Geographical Information Systems (GIS) are a new technology that are often used in applications requiring analysis and manipulation of the spatial conditions and situations. It provides spatial and non-spatial databases, as well as various analytical tools for general application purposes. The characteristics of a GIS are such that it can be used for radio signal path-loss prediction. This thesis examines the existing analytical path-loss prediction methods. The possibility and feasibility of using GIS analytical tools to predict signal propagation path-loss is discussed. GIS functions, using TIN as a terrain representation are used to carry out qualitative and quantitative analysis of signal propagation based on the analytical prediction models are examined and prototyped. The method is implemented using AML programs in ARC/INFO software, and C functions on a SUN Spare workstation. The test is performed on the Hobart area in Tasmania, Australia. Program performance speed and accuracy are analyzed and discussed against the method based on grid terrain model.