A robust prediction model using ANFIS based on recent TETRA outdoor RF measurements conducted in Riyadh city – Saudi Arabia

TLDR: Adapt neuro-fuzzy inference system (ANFIS) is used as a robust wireless signal predictor and it turns out that the ANFIS prediction model outperforms the predictors based on empirical models, and is marginally better than RBF-NN predictor.

Abstract: Received wireless signal prediction is a difficult and complex task. Various types of prediction models such as deterministic, empirical, as well as statistic were developed. However, they rarely adapt well to different types of environments. Prediction models based on artificial intelligence techniques are the recent alternative approaches to predict the signal strength at a particular location in an investigated area. The advantage of using artificial intelligence for field strength prediction is given by the flexibility to adapt to different environments, high-speed processing, and the ability to process a high quantity of data. In this paper, adaptive neuro-fuzzy inference system (ANFIS) is used as a robust wireless signal predictor. The performance of the proposed predictor is then compared to the predictors based on radial basis function neural network (RBF-NN), and three most widely used empirical path loss models. The performance criterion selected for the comparison between the actual and the predicted data are the root mean square error (RMSE), maximum relative error (MRE), and goodness of fit ( R 2 ). It turns out that the ANFIS prediction model outperforms the predictors based on empirical models, and is marginally better than RBF-NN predictor.