Modelling Large-Scale Signal Fading in Urban Environment Based on Fuzzy Inference System

Path loss models are veritable tools for estimation of expected large-scale signal fading in a specific propagation environment during wireless network design and optimization. In this paper, the capability of Adaptive Neuro-Fuzzy Inference System (ANFIS) to establish non-linear relationship between related variables was explored for path loss predictions at Very High Frequency (VHF) band in a typical urban propagation environment. Drive test measurements were conducted along various routes in the urban area to obtain terrain profile data and path losses of radio signals transmitted at 92.3 MHz and 189.25 MHz frequencies. ANFIS was modelled to predict the magnitude of large-scale signal fading (i.e. path loss) based on the longitude, latitude, distance and elevation of the receiver’s location. Fuzzy Inference System (FIS) was generated based on Fuzzy C-Means (FCM) and subtractive clustering methods. Model performance evaluation results showed that the ANFIS model developed based on FCM clustering method yielded the least prediction errors with a Root Mean Squared Error (RMSE) value of 0.88 dB. Whereas, the International Telecommunications Union Radiocommunication (ITU-R) had earlier set a maximum allowable RMSE value of 6 dB for urban propagation environments. Thus, ANFIS technique produced a very efficient largescale signal fading prediction model for VHF network design and optimization in urban areas.