Robust mobile geo-location algorithm based on LS-SVM

Author

Sun, Guolin ; Guo, Wei

Author_Institution

Sch. of Commun. & Inf. Eng., Univ. of Electron. Sci. & Technol. of China, Chengdu, China

Volume

54

Issue

3

fYear

2005

fDate

5/1/2005 12:00:00 AM

Firstpage

1037

Lastpage

1041

Abstract

Support vector machine (SVM) is powerful to solve problems such as nonlinear classification, function estimation and density estimation. It has also led to many other recent developments in kernel-based learning fields. In this paper, we extend a high-accuracy, real-time, and fault-tolerant SVM to mobile geo-location problem, which has become an important component of pervasive computing. Simulation results show its basic location performance superior to conventional least square (LS) algorithm especially under nonlight of sight (NLOS) environments. Finally, we also analyze the impacts of training samples and training area on test location accuracy.

Keywords

cellular radio; fault tolerance; least squares approximations; radionavigation; support vector machines; ubiquitous computing; density estimation; fault-tolerant SVM; function estimation; kernel-based learning fields; least square algorithm; nonlight of sight; nonlinear classification; robust mobile geo-location algorithm; support vector machine; Base stations; Cellular networks; FCC; Geometry; Mobile computing; Neural networks; Robustness; Support vector machine classification; Support vector machines; Wireless networks; Least square (LS) support vector machine (SVM); mobile geo-location; nonlight of sight (NLOS);

fLanguage

English

Journal_Title

Vehicular Technology, IEEE Transactions on

Publisher

ieee

ISSN

0018-9545

Type

jour

DOI

10.1109/TVT.2005.844676

Filename

1433248