Paper Info
Title
Investigating Geometry-Aware Network-Based Positioning in Cellular Networks Using Neural Network Predictive Model.
Abstract
Received signal strength (RSS) based positioning techniques are one of the easy and cost effective methods of estimating the location of a mobile station with no Global Positioning System. The major challenge with received signal strength based positioning is high variability in its accuracy due to varying influences in its relationship with parameters such as wireless environment, geometry of the network, propagation model, change in infrastructure and so on. This paper addresses the influence geometry of base stations where location estimation measurements are obtained has on the accuracy of three received signal strength based geometric techniques. A neural network (NN) model for the geometry-aware cellular network was developed using input---output dataset obtained from the system's model. Levenberg---Marquardt algorithm was employed in training the NN model developed. Results obtained indicated that the geometry of the cellular network used for the network-based positioning had significant impact on the accuracy of the location estimation. Comparison of the influence of network geometry between RSS-based Centre of Gravity, Circular Trilateration and Least Square algorithms using the neural network model were discussed and a critical analysis of the result presented.
Year
DOI
Venue
2016
10.1007/s11277-016-3401-y
Wireless Personal Communications
Keywords
Field
DocType
RSS,LBS,Network-based positioning,Location accuracy,Network geometry,Geometry-aware
Base station,Wireless,Computer science,Mobile station,Real-time computing,Cellular network,Global Positioning System,Geometry,Artificial neural network,RSS,Trilateration
Journal
Volume
Issue
ISSN
90
3
0929-6212
Citations 
PageRank 
References 
0
0.34
2
Authors
3
Name
Order
Citations
PageRank
Folasade Mojisola Dahunsi100.34
Dwolatzky, B.2112.33
DahunsiFolasade Mojisola300.34