MULTI-LAYER PERCEPTRON NEURAL NETWORK UHF FIELD STRENGTH PREDICTION MODEL FOR MAIDUGURI METROPOLIS

Abstract

This paper investigates the application of a Multi-layer Perceptron Neural Network (MLP-NN) based model for field strength prediction across the Maiduguri metropolis at an operating frequency of 1800MHz. Received power values obtained from multiple Base Transceiver Stations situated within the city were used to train, validate and test the MLP-NN for ability to generalize. Results indicate that the MLP-NN model with a Root Mean Squared Error (RMSE) value of 5.29dB offers an improvement over the COST 231 Walfisch-Ikegami model, which has an RMSE value of 7.95dB.