Adaptive propagation modelling based on neural network techniques

A hybrid modelling approach for the prediction of terrestrial wave propagation is introduced. It is shown that the use of neural networks leads to highly adaptive models by the simultaneous representation of nonlinear dependencies of various environmental parameters. This flexible and computationally effective approach can be used for calibration as well as extension of conventional prediction models. Measurement results are presented in urban microcells, showing that this idea leads to significant improvements in practical planning scenarios