UTAR Institutional Repository

An Accurate and Efficient Shooting-and-bouncing-polygon Ray Tracer For Radio Propagation Modelling

Teh, Chin Hui (2019) An Accurate and Efficient Shooting-and-bouncing-polygon Ray Tracer For Radio Propagation Modelling. PhD thesis, UTAR.

Download (9Mb) | Preview


    Radio propagation modelling is important for the design and deployment of wireless communication systems. Ray tracing is the state-of-the-art sitespecific technique to modelling terrestrial and indoor propagation. In this project, we have proposed and implemented three important improvements to a 3D shooting-and-bouncing-polygon (SBP) ray tracer. Firstly, we have derived delay correction factors to be used with a onepatch model of real building wall with thickness. The correction factors have been derived for reflection, transmission, and diffraction, based on a multilayer lossy wall model. The one-patch model and correction factors allow efficient and accurate treatment of real building walls with thickness in ray tracing. Secondly, we have extended SBP to trace diffracted and diffractedreflected ray beams using edge-fixed diffraction ray-polygons. Crucial to the solution is the ability to transform or project Cartesian polygons to edge-fixed polygons. We have described the associated problems and proposed an algorithm to perform the transformation or projection.Thirdly, we have proposed a new spatial partitioning scheme to be used with the SBP ray tracer. The new scheme is named convex cell partitioning (CCP). It divides the simulated scene into a set of interconnected convex cells of any shapes. The interconnectivity between the convex cells is represented by a graph. Like binary-space-partitioning, CCP is able to sort objects by their visibility distance relative to any point, efficiently, incrementally, and adaptively. Better than binary-space-partitioning, in the context of SBP, CCP removes the need to perform polygon subtraction, a relatively expensive procedure. Traversal of the CCP graph is also simpler than the binary tree. To improve CCP usability, we have developed and implemented a tool to automatically construct the CCP graph from a given set of 3D walls. The improved ray tracer is named SBP-CCP. We have evaluated its performance by making comparisons with commercial ray tracers, full-wave solutions, and published measurements. Results show that SBP-CCP is more accurate than the earlier version and it is about an order of magnitude faster than the binary-space-partitioning version in an urban canyon application. It is also shown that SBP-CCP outperforms commercial REMCOM Wireless InSite 3D ray tracers in terms of accuracy and time and memory efficiency in long tunnel, urban canyon, and indoor applications. In particular, SBP-CCP has very small memory footprint, 2 to 3 orders of magnitude smaller than REMCOM ray tracers. SBP-CCP simulation results also show very good match (2 dB rootmean-square error) to full-wave solutions computed using CST Microwave.Studio time domain solver, in an indoor application with reflection, transmission, diffraction, and lossy walls with thickness. SBP-CCP is full 3D, fast, accurate, and memory efficient. It is a good candidate for simulating long tunnel, urban canyon, and indoor propagation environments. The current implementation does not include double diffraction and it is not suitable for urban environments with multiple over-roof-top diffractions. The current implementation is also not suitable for applications where diffuse scattering is an important propagation mechanism.

    Item Type: Final Year Project / Dissertation / Thesis (PhD thesis)
    Subjects: T Technology > TK Electrical engineering. Electronics Nuclear engineering
    Divisions: Institute of Postgraduate Studies & Research > Lee Kong Chian Faculty of Engineering and Science (LKCFES) - Sg. Long Campus > Doctor of Philosophy in Engineering
    Depositing User: Sg Long Library
    Date Deposited: 04 Dec 2019 17:45
    Last Modified: 04 Dec 2019 17:45
    URI: http://eprints.utar.edu.my/id/eprint/3603

    Actions (login required)

    View Item