Application of an automatic tool for the planning of a cellular network in a real town

An application of an automatic tool for the planning of a cellular GSM network in a real environment is presented. The basic principles of the algorithm were previously developed by the authors, and in this paper solutions to the problems arising from its application to a real town are proposed. One of the main issues concerns the field prediction models, so two algorithms are used in the real environment: one is based on an artificial neural network (ANN), and one on the Cost231 model, modified for hilly terrain, and in both cases the position and the height of the buildings are considered in detail. The whole planning procedure is applied to the town of Ancona and the results give the optimized location of the radio base stations (RBS), the heights of their antennas, and their transmitted power.