Radio wave propagation prediction modelling has become increasingly important in wireless network planning and optimization since the emergence of 3G networks. The propagation predictions serve as a fundamental output for the advanced analysis and optimization such as capacity and link budgets etc. The identification of QoS (Quality of Services) or weak signal spots are based on the estimation of propagation prediction. At present, in order to build an efficient indoor DAS (Distributed Antenna System), the candidate antennas need to be tested and the possible combinations evaluated to assess which will give an optimal solution for indoor antenna placement. The solution is based on propagation prediction modelling because the minimal coverage ratio has to be considered.
Currently, radio wave propagation models consist of two types: small-scale and large-scale.
• Small-scale propagation prediction deals with fast fading (i.e. the variation of signal strength over a short period of time such as one wavelength).
• Large-scale propagation prediction computes the average signal strength over a longer period of time.
The modelling of large-scale radio wave propagation in indoor environments plays a crucial role in the investigation of 3G/4G network planning applications (such as localisation). In indoor environments, there are usually more irregular objects and material types, which make modelling much more complex when compared to outdoor environments.
In general, large-scale propagation models fall into two categories: empirical and deterministic. Empirical models are mainly based on empirical factors such as distance or frequency. They are computationally fast but they do not consider a great deal of environmental information so their accuracy is limited. Deterministic approaches take into account the environmental information such as object positions and the corresponding materials. Generally speaking, these approaches are more time-consuming when compared to empirical models but allow for a higher level of accuracy to be obtained. Despite many acceleration techniques being applied, the use of accurate propagation modelling for indoor scenarios remains limited due to the complex indoor propagation environment. Aside from these two categories, some propagation models consider both empirical and deterministic factors, which are categorized as semi-empirical (or semi-deterministic) approaches. Such models usually perform faster than deterministic approaches such as ray tracing and their accuracy is high in some scenarios. Ray-based methods can be categorized as deterministic approaches. They are widely used in propagation prediction. Compared to FDTD (Finite Difference Time Domain)-like methods, they consume less memory and are far more efficient. These ray-based methods compute the possible rays between the emitter and receivers in complex environments and they need to search the rays to compute reflections and diffractions, based on Descartes’ laws. Hence, they tend still to be very time-consuming if the environment is complex, i.e., if there are a large number of obstacles. Usually, the accuracy of ray-based methods is limited by the number of rays that can be computed within a reasonable time.