Novel Approach to Indoor Electromagnetic Wave Propagation Modelling

In the 1980s, Prof. Novotný and Dr. Pechač put heads together at the Faculty of Electrical engineering at the Czech Technical University in Prague, and thought up first fundaments of easy and rapid method of ray propagation. After graduating, I started enthusiastically working on indoor propagation modelling based on semi-empirical models and on full ray tracing prediction within an anechoic shielding chamber [81-93]. Semi-empirical models are easily understandable and applicable, which is their big advantage is. However, they did not satisfy my expectation that they would provide a correct and clear prediction of signal coverage. Full ray tracing is very potent, but it is too rigorous and too complicated for a complex indoor environment. Neither semi-empirical approach nor full ray-tracing approach seemed to me to be well suited for indoor propagation. For this reason I started to work on something new, which seemed to be more promising for indoor propagation prediction. First, I began to implement a ray tracing model. This was a really complicated task, and it led me to reflect on the special features of an indoor environment. New ideas began to take root. This resulted in the suggestion of ray propagation within a grid and precalculated local environment behaviour. During a consultation with my supervisor Dr. Pechač, he retrieved old ideas of a method of easy and rapid ray propagation that had never been checked out. He also recommended implementing them into new model. This was a crucial suggestion, for which I am grateful. The idea was, firstly, to make a probabilistic description of a local environment including a probabilistic radiation pattern of the obstacles and, secondly, to represent the signal level by simple ray density. These two thoughts, together with grid propagation, remain the principal elements of the current model. These new, stimulating recommendations led a well-working implementation of a new model into a software package and to the first publications of description if the new model, at that time known as the XY-model [94-96]. The first success was followed by applications of the model not only for signal coverage prediction, but also for prediction of some wideband channel parameters. This was accompanied by several publications [101-103, 108, and 109] and the name of the model was changed to the Motif Model. A further great stimulus while working on the new model was provided during my fellowship in Ireland by Dr. Pesch, who was also captivated …