Hybrid Field Strength Prediction Model With Ray Tracing Based On Dominant Path And Artificial Neural Networks

In micro-cell and pico-cell environment,the statistical similarity does not exist in the different cells.Using statistical models to predict filed strength has a greater error.And to obtain accurate results in deterministic model is on the premise that it has high precision scene database and uses highly complex algorithm which results in a large amount of computation and low efficiency.This paper introduces the artificial neural network because of the above limitations of deterministic model for predicting field strength and presents hybrid prediction model for field strength based on ray tracing method with dominant path and artificial neural networks.This paper presents a method for extracting the dominant path of ray tracing and defines the dominant path as the direct path or the primary reflection path.First,using ray tracing tool based on the dominant path to get the filed strength prediction data and the information of the dominant path.Then,get the measured data by using the USRP device for measuring indoor scenes.Predicted data is obtained by training RBF neural network through a part of actual measured values and information of dominant path.To verify the feasibility of the hybrid prediction model in indoor environment,this paper compares the predicted data and measured data.Compared with ray tracing method,the hybrid model increases prediction accuracy;and also this hybrid model uses ray tracing method’s dominant path to extract more information to close to the real scenario which decreases complexity.