Source Localization in Urban Environments Using Realistic 3-D Models and Ray Tracing Simulation
Date
2016-05
Authors
Contributor
Advisor
Department
Instructor
Depositor
Speaker
Researcher
Consultant
Interviewer
Narrator
Transcriber
Annotator
Journal Title
Journal ISSN
Volume Title
Publisher
University of Hawaii at Manoa
Volume
Number/Issue
Starting Page
Ending Page
Alternative Title
Abstract
The focus of this thesis is the exploration and development of source localization methods which are effective in urban environments. Multipath, non-line-of-sight, and other effects which dominate in complex environments can be challenging for traditional localization techniques. To investigate source localization in realistic urban environments, 3-D geographical models of real-world areas are reconstructed from openly available geospatial resources. Ray tracing is suitable for large and complex geometries, and is employed along with these constructed models to realistically model radio propagation in such environments. Two source localization methods are presented, including time reversal analysis using ray tracing as well as a machine learning approach. The time reversal approach proves to benefit from multipath conditions. The focusing effect of the time reversal process can achieve sub-wavelength localization accuracies. However, the time reversal area observation technique and the resulting computational requirement limits the practical size of the area in which this method can be applied. In contrast, the machine learning approach shows to be accurate and efficient even for large areas. Machine learning, using example data generated for specific locations, can provide a best-effort prediction on the general vicinity of a transmitting source. However, the achievable localization accuracy is reduced compared to the time reversal approach. A localization scheme combining these methods is demonstrated, showing that the techniques are complimentary. The machine learning method can efficiently provide an initial localization estimate, significantly reducing the area required in the time reversal step. The time reversal localization method can then be used around the initial machine learning prediction, providing a verification of the presence of the source, as well as further increasing the accuracy of the localization result. In one presented case, the machine learning estimate predicts the source location to within 0.69 m of its true location, and the time reversal method further improves the result to within 0.05 m of the source’s true location.
Description
Keywords
Citation
Extent
Format
Geographic Location
Time Period
Related To
Theses for the degree of Master of Science (University of Hawaii at Manoa). Electrical Engineering
Related To (URI)
Table of Contents
Rights
Rights Holder
Local Contexts
Collections
Email libraryada-l@lists.hawaii.edu if you need this content in ADA-compliant format.