Non-linear, non-gaussian estimation for INS/GPS integration

Thanh Trung Duong, Kai-Wei Chiang

Research output: Contribution to journalArticle

1 Citation (Scopus)

Abstract

In a mobile mapping system, the integration of Inertial Navigation System and Global Positioning System is widely applied for determining position and orientation. The Kalman filter or Extended Kalman filter is popularly used for multi-sensor fusion estimation. In those estimation strategies, the linearization and assumption of Gaussian distribution are utilized. However, the fact that the state and measurement models in the integration of Inertial Navigation System and Global Positioning System are originally non-linear and the noise arising during operation may be non-Gaussian distribution. This characteristic may lead to the degraded performance of the system utilizing Kalman filter or Extended Kalman filter in case of non-linear model and non-Gaussian noises. This paper investigates some of non-linear, non-Gaussian estimation strategies in order to improve the performance of an integrated system. Firstly, the fundamental of algorithms will be introduced, those estimation strategies are then applied on simulation scenario to find the optimal method, finally, real measurements collected from field test will be applied to verify the performance of proposed algorithms.

Original languageEnglish
Pages (from-to)1081-1086
Number of pages6
JournalSensor Letters
Volume10
Issue number5-6
DOIs
Publication statusPublished - 2012 May 1

Fingerprint

Kalman filters
Global positioning system
inertial navigation
Inertial navigation systems
Global Positioning System
Extended Kalman filters
Sensor data fusion
multisensor fusion
field tests
Gaussian distribution
linearization
normal density functions
Linearization
simulation

All Science Journal Classification (ASJC) codes

  • Atomic and Molecular Physics, and Optics
  • Electrical and Electronic Engineering

Cite this

Duong, Thanh Trung ; Chiang, Kai-Wei. / Non-linear, non-gaussian estimation for INS/GPS integration. In: Sensor Letters. 2012 ; Vol. 10, No. 5-6. pp. 1081-1086.
@article{c5a5a92f242b4c2e93df10aebdaba438,
title = "Non-linear, non-gaussian estimation for INS/GPS integration",
abstract = "In a mobile mapping system, the integration of Inertial Navigation System and Global Positioning System is widely applied for determining position and orientation. The Kalman filter or Extended Kalman filter is popularly used for multi-sensor fusion estimation. In those estimation strategies, the linearization and assumption of Gaussian distribution are utilized. However, the fact that the state and measurement models in the integration of Inertial Navigation System and Global Positioning System are originally non-linear and the noise arising during operation may be non-Gaussian distribution. This characteristic may lead to the degraded performance of the system utilizing Kalman filter or Extended Kalman filter in case of non-linear model and non-Gaussian noises. This paper investigates some of non-linear, non-Gaussian estimation strategies in order to improve the performance of an integrated system. Firstly, the fundamental of algorithms will be introduced, those estimation strategies are then applied on simulation scenario to find the optimal method, finally, real measurements collected from field test will be applied to verify the performance of proposed algorithms.",
author = "Duong, {Thanh Trung} and Kai-Wei Chiang",
year = "2012",
month = "5",
day = "1",
doi = "10.1166/sl.2012.2267",
language = "English",
volume = "10",
pages = "1081--1086",
journal = "Sensor Letters",
issn = "1546-198X",
publisher = "American Scientific Publishers",
number = "5-6",

}

Non-linear, non-gaussian estimation for INS/GPS integration. / Duong, Thanh Trung; Chiang, Kai-Wei.

In: Sensor Letters, Vol. 10, No. 5-6, 01.05.2012, p. 1081-1086.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Non-linear, non-gaussian estimation for INS/GPS integration

AU - Duong, Thanh Trung

AU - Chiang, Kai-Wei

PY - 2012/5/1

Y1 - 2012/5/1

N2 - In a mobile mapping system, the integration of Inertial Navigation System and Global Positioning System is widely applied for determining position and orientation. The Kalman filter or Extended Kalman filter is popularly used for multi-sensor fusion estimation. In those estimation strategies, the linearization and assumption of Gaussian distribution are utilized. However, the fact that the state and measurement models in the integration of Inertial Navigation System and Global Positioning System are originally non-linear and the noise arising during operation may be non-Gaussian distribution. This characteristic may lead to the degraded performance of the system utilizing Kalman filter or Extended Kalman filter in case of non-linear model and non-Gaussian noises. This paper investigates some of non-linear, non-Gaussian estimation strategies in order to improve the performance of an integrated system. Firstly, the fundamental of algorithms will be introduced, those estimation strategies are then applied on simulation scenario to find the optimal method, finally, real measurements collected from field test will be applied to verify the performance of proposed algorithms.

AB - In a mobile mapping system, the integration of Inertial Navigation System and Global Positioning System is widely applied for determining position and orientation. The Kalman filter or Extended Kalman filter is popularly used for multi-sensor fusion estimation. In those estimation strategies, the linearization and assumption of Gaussian distribution are utilized. However, the fact that the state and measurement models in the integration of Inertial Navigation System and Global Positioning System are originally non-linear and the noise arising during operation may be non-Gaussian distribution. This characteristic may lead to the degraded performance of the system utilizing Kalman filter or Extended Kalman filter in case of non-linear model and non-Gaussian noises. This paper investigates some of non-linear, non-Gaussian estimation strategies in order to improve the performance of an integrated system. Firstly, the fundamental of algorithms will be introduced, those estimation strategies are then applied on simulation scenario to find the optimal method, finally, real measurements collected from field test will be applied to verify the performance of proposed algorithms.

UR - http://www.scopus.com/inward/record.url?scp=84866237502&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84866237502&partnerID=8YFLogxK

U2 - 10.1166/sl.2012.2267

DO - 10.1166/sl.2012.2267

M3 - Article

VL - 10

SP - 1081

EP - 1086

JO - Sensor Letters

JF - Sensor Letters

SN - 1546-198X

IS - 5-6

ER -