dc.description.abstract
The characterization of the wireless propagation channel provides the fundamental basis for wireless communications, determining the performance of practical systems. In last decades, the dramatically increasing demand for wireless system capabilities makes the study of the propagation channel become indispensable. An accurate and low-complexity channel model is of high importance for developing next generation wireless systems. Until now, there are mainly three modeling methods: empirical, statistical and deterministic models. The latter uses the geometric and electromagnetic description for a site-specific environment to evaluate the propagation paths based on geometrical optic techniques. Ray tracing (RT), a deterministic propa- gation prediction tool, has been widely used to simulate channel characteristics in indoor and outdoor environments. To date, RT tools include not only specular reflection, penetration through dielectric blocks and diffraction, but also diffuse scattering mechanisms. The accuracy, provided by a detailed modeling of the environment, comes at the cost of a high computational complexity, which directly scales with the number of considered propagation paths. The goal of my thesis is to reduce the computational complexity of RT with no loss in accuracy. There are three scenarios included: wideband indoor, ultra-wideband (UWB) indoor and tunnel scenarios. A three-dimensional (3D) RT tool used in the thesis is based on the pre-existent RT tool, which was implemented in C programming language by Universite catholique de Louvain, Belgium. Firstly, the RT tool is re-implemented in MATLAB, which is named conventional RT in the thesis. For the sake of accelerating the execution of the RT tool, the code is optimized through converting time-consuming algorithms to Matlab executable (MEX) functions by using MATLAB Coder. The speeding up efforts focus on reflection and diffuse scattering calculations, because the number of reflection and diffuse scattering propagation paths comprise a large proportion among all propagation paths. Compared with conventional RT, the simulation time of the revised MATLAB code is significantly reduced. Moreover, the reduction of computational complexity of RT is considered not only for one terminal position but also for multiple mobile terminal positions in this thesis. For one terminal position, an efficient approach to generate diffuse scattering tiles based on concentric circles is developed and evaluated for a wideband indoor scenario. It is known that channel characteristics may vary significantly over the entire bandwidth for an UWB indoor scenario. To cope with this, sub-band divided ray tracing (SDRT) has been proposed for one terminal position. However, the compu- tational complexity is directly proportional to the number of sub-bands. Therefore, I propose a mathematical method by making SDRT almost independent of the number of sub-bands, which is named low-complexity SDRT. In addition, RT combining with a higher-order reflection algorithm is developed for intelligent transport system (ITS) in tunnel scenarios. While simulating the radio propagation conditions for a mobile terminal, communicating in a frame based communication system indoors with several fixed nodes, the correlated temporal and spatial evolution of the channel impulse response (CIR) is of utmost concern. A RT algorithm and a low-complexity SDRT algorithm based on two-dimensional discrete prolate spheroidal (DPS) sequences is proposed for wide-band and UWB indoor scenarios, respectively. In tunnel scenarios, a non-stationary vehicle-to-vehicle (V2V) channel model combining propagation graphs with RT is proposed. I include the time evolution of relevant channel parameters in the pro- posed model depending on the stationary time region, which are obtained based on the local scattering function (LSF). Furthermore, the accuracy of RT is strictly limited by the available description of the environment. Based on the low-complexity SDRT implementation, I also propose a calibration method for indoor UWB low-complexity SDRT. The method estimates the optimal material parameters, including the dielectric parameters and the scattering parameters, using channel measurements and multiobjective simulated annealing (MOSA). Finally, the accuracy of all proposed algorithms is verified by numeric simulations or measurement campaigns. The evaluation includes the power delay profile (PDP), root mean square (RMS) delay spread, angular spread, and RMS Doppler spread. Due to the complicated implementation of RT, it is hard to obtain a closed-form expression for the computational complexity. Therefore, I evaluate the simulation time of my proposed algorithms. In conclusion, all the proposed algorithms in this thesis can help to reduce the computational complexity of RT significantly.
en
