RADIO PROPAGATION MODELING OF JOS SUBURBS AT 900MHZ USING AN ADAPTIVE NEURO-FUZZY INFERENCE SYSTEM

Résumé

This study investigates radio propagation modeling of the suburbs of Jos, Nigeria, at an operating frequency of 900MHz, using an Adaptive Neuro-Fuzzy Inference System (ANFIS) Technique. Path loss values computed based on received power measurements obtained from Base Transceiver Stations situated across the suburbs of the city were used to train, validate and test the ANFIS model for capacity to predict path loss. Results indicated that the ANFIS model with a Root Mean Square Error (RMSE) value of 4.74dB offers an improvement in prediction accuracy over the COST 231 Hata model, which has an RMSE value of 5.90dB.