Modelling and coverage improvement of DVB-T networks

Abstract

The necessity of accurate point-to-area and point-to-point prediction tools arises from the enormous demand in designing broadcasting systems for digital TV and cellular communications. Up to now, a considerable number of coverage prediction models for radio coverage has been developed. In electromagnetic wave propagation theory, there are three types of propagation models. Empirical models that are based on a large quantity of measurement data are elementary but not very accurate. Semi-deterministic models that are based on measurement data and electromagnetic theory of propagation, which are more precise. Finally, deterministic models based on theoretical physics, like diffraction theory and Fresnel theory, that require a significant amount of geometrical data about the propagation terrain profile but are the most accurate. The primary outcomes of this research are the comparative study and improvement of several propagation models, using a significant quantity of measurements and simulations and the deduction of useful conclusions to be used by engineers to improve propagation predictions further. In this research, the Longley-Rice (ITM) Irregular Terrain Model model was used, a classic model used for TV coverage prediction, which model is to date the preferred model of the FCC (Federal Communications Commission) in the US for FM-TV coverage calculations. To run the model, the Radio Mobile program (Radio Propagation and Virtual Mapping Freeware) was used based on the Longley-Rice Model ITM, including the 3-arc-second Satellite Radar Terrain Mission (SRTM) maps and the SPLAT! program (an RF Signal Propagation, Loss, And Terrain analysis tool), which also relies on the Longley-Rice ITM model and makes use of SRTM maps. Both programs work in Windows operating system (Windows7 Professional, 64 bits). Another model used in this research was SPLAT! with ITWOM (Irregular Terrain with Obstructions Model) which combines empirical data from the ITU-R P.1546 model and other ITU recommendations in conjunction with Beer's and Snell's laws. The ITU-R Recommendation P.1546 model and the empirical Hata-Davidson model using HAAT were also utilized in this research. The Single Knife-Edge (SKE) model was coded in MATLAB and utilized in this research as a simple reference model, where only one main obstacle is considered. Other well-known multiple knife-edge diffraction models employed in this study are the Epstein-Peterson, Deygout, and Giovaneli models. For these deterministic models, individual MATLAB programs were written. Simulations produced by the models were limited to the main two knife-edges of the propagation path for immediate comparison with the Longley-Rice model which uses the “double knife-edge” approach. All measurement campaigns took place in Northern Greece and Southern (F.Y.R.O.M) Former Yugoslav Republic of Macedonia using a Rohde & Schwarz FSH-3 portable spectrum analyser and precision calibrated antennas.