Investigation of Time-Difference-Of-Arrival Localization Method for Non-Line-Of-Sight Scenarios

Date

2016-12

Journal Title

Journal ISSN

Volume Title

Publisher

Abstract

A novel localization concept, which considers non-line-of-sight (NLOS) propagation, is proposed in this dissertation. By introducing a transfer function that relates the field at a given receiver to the source as a function of frequency and position, the NLOS effects can be mitigated and the propagation channel can be calibrated back to free space. The conventional Time Difference of Arrival (TDOA) method under Line-of-Sight (LOS) and noisy conditions is implemented first, and then the theory is extended to the proposed method, which de-embeds the transfer function and extends the TDOA method to account for NLOS effects. A global search method and an iterative method are both introduced, based on the concept of de-embedding the transfer function. This involves using processed received signals (received signals after de-embedding the corresponding transfer functions). The localization accuracy achieved using the iterative method for various NLOS scenarios is investigated, including a refraction problem, a near field problem, and a scattering problem. Subsequently, the study is further extended to allow for multiple reflection effects. Three different models are provided. First, a parallel-wall model considers the case when the reflections are basically coming from a single horizontal direction. A 3D localization model and 2D enclosed wall model are then also presented to illustrate cases when the reflections come from either the vertical or the two horizontal directions. A hybrid iterative method, combining the iterative method and the global search method, is proposed to solve the problem of convergence that occurs when the transmitted signal is operating at high frequencies (so that the enclosed area of interest is large compared with a wavelength), especially with a narrow frequency bandwidth. Simulation results demonstrate that the proposed algorithm performs much better than the usual TDOA localization method, and the accuracy is only limited by the sampling frequency. An experiment (performed at Sandia National Laboratories) is set up to verify the improvement in using the proposed method vs. the usual TDOA method.

Description

Keywords

Localization, Time-difference-of-arrival, Electromagnetics, Iterative method, Global search method, Refraction, Scattering, Reflection

Citation