Comparative Analysis of Ray tracing, finite integration technique and empirical models using ultra-detailed indoor environment model and measurements

Abstract

3D Finite Integration Technique (FIT), ray tracing and simple empirical model accuracy is compared on a single basis, using a high-detailed environment model. The most accurate FIT Maxwell's equations solver demonstrated very high precision (1.1 – 2.4 dB) when compared with measurements in a complex indoor environment, thus validating the environment model and measurement campaign for ray tracing accuracy estimation. The comparison of methods has been made generally in terms of accuracy of narrowband (field level spatial distributions) and wideband (power delay profiles) channel characteristics prediction as well as required computing resources. Another contribution of current work is the investigation of the prediction accuracy dependence on the spatial detailing of prediction (size of spatial averaging window as an example). It is shown that prediction accuracy increases significantly with the decrease in spatial detailing of prediction, though in some cases the increase in accuracy is not monotonic. The impact of the environment model detail and its spatial size on the accuracy of predictions is also investigated. The environment model included double-framed windows, doors, metallic objects, some meaningful furniture, along with precision geometry.