Queen's University - Utility Bar

QSpace at Queen's University >
Graduate Theses, Dissertations and Projects >
Queen's Graduate Theses and Dissertations >

Please use this identifier to cite or link to this item: http://hdl.handle.net/1974/5945

Title: Advanced Nonlinear Techniques for Low Cost Land Vehicle Navigation
Authors: Georgy, Jacques

Files in This Item:

File Description SizeFormat
Georgy_Jacques_F_W_201007_PhD.pdf13.86 MBAdobe PDFView/Open
Keywords: Land Vehicle Navigation
Global Positioning System (GPS)
Inertial Navigation System (INS)
MEMS-based INS/GPS Integration
Particle Filter
Kalman Filter
Issue Date: 2010
Series/Report no.: Canadian theses
Abstract: Present land vehicle positioning and navigation relies mostly on the Global Positioning System (GPS). However, in urban canyons, tunnels, and other GPS-denied environments, the GPS positioning solution may be interrupted or suffer from deterioration in accuracy due to satellite signal blockage, poor satellite geometry or multipath effects. In order to achieve continuous positioning services, GPS is augmented with complementary systems capable of providing additional sources of positioning information, like inertial navigation systems (INS). Kalman filtering (KF) is traditionally used to provide integration of both INS and GPS utilizing linearized dynamic system and measurement models. Targeting low cost solution for land vehicles, Micro-Electro-Mechanical Systems (MEMS) based inertial sensors are used. Due to the inherent errors of MEMS inertial sensors and their stochastic nature, which is difficult to model, KF has limited capabilities in providing accurate positioning in challenging GPS environments. This research aims at developing reliable integrated navigation system capable of demonstrating accurate positioning during long periods of challenging GPS environments. Towards achieving this goal, Mixture Particle filtering (MPF) is suggested in this research as a nonlinear filtering technique for INS/GPS integration to accommodate arbitrary inertial sensor characteristics, motion dynamics and noise distributions. Since PF can accommodate nonlinear models, this research develops total-state nonlinear system and measurement models without any linearization, thus enabling reliable integrated navigation and mitigating one of the major drawbacks of KF. Exploiting the capabilities of PF, Parallel Cascade Identification (PCI), which is a nonlinear system identification technique, is used to obtain efficient stochastic models for inertial sensors instead of the currently utilized linear models, which are not adequate for MEMS-based sensors. Moreover, this research proposes a method to update the stochastic bias drift of inertial sensors from GPS data when the GPS signal is adequately received. Furthermore, a technique for automatic detection of GPS degraded performance is developed and led to improving the performance in urban canyons. The performance is examined using several road test experiments conducted in downtown cores to verify the adequacy and the benefits of the methods suggested. The results obtained demonstrate the superior performance of the proposed methods over conventional techniques.
Description: Thesis (Ph.D, Electrical & Computer Engineering) -- Queen's University, 2010-07-23 20:27:02.12
URI: http://hdl.handle.net/1974/5945
Appears in Collections:Queen's Graduate Theses and Dissertations
Department of Electrical and Computer Engineering Graduate Theses

Items in QSpace are protected by copyright, with all rights reserved, unless otherwise indicated.


  DSpace Software Copyright © 2002-2008  The DSpace Foundation - TOP