Modeling by optimal Artificial Neural Networks the prediction of propagation path loss in urban environments

Abstract

In this paper we present an alternative procedure for the prediction of propagation path loss in urban environments, which is based on Artificial Neural Networks (ANN). The goal of this work is to synthesize and model ANNs which would require entering at the input nodes a detailed and the same time small amount of information about the propagation environment. We apply the Differential Evolution (DE) algorithm, in conjunction with the Levenberg-Marquardt backpropagation algorithm in order to train different ANNs. The combined DE-LM method achieves better convergence of neural network weight optimization. We present two different ANN design cases with different number of input nodes. The general performance of the both ANNs shows their effectiveness to yield results with satisfactory accuracy in short time. The received results are compared to the respective ones yielded by the Ray-Tracing model and exhibit satisfactory accuracy.