Paper Info
Title
Smartphone based indoor localization using stable access points.
Abstract
Indoor localization, based on Wi-Fi signals, is becoming a popular approach for providing location based services in indoor environment. The challenging task of accurately finding the position of a device depends on prior efforts of fingerprinting. However, fingerprint data are susceptible to many indoor environmental factors such as, change of furniture or conditions like door and window open/close. Thus it is important to detect a minimal set of stable Access Points ( APs ) for providing durable localization services in different training and test condition. Consequently, in this paper, we propose radio frequency based indoor localization technique that selects stable APs to achieve appreciable accuracy, irrespective of device, temporal or context heterogeneity (door and window open/close, presence/absence of other users in vicinity). Hence, we have created an indoor localization data set where data is collected from entire floor of a building in our University campus. Our proposed algorithm returns minimum set of stable APs to cover each location point of the experimental region with at least 3 APs. With supervised learning techniques, the system is found to yield 84.60% localization accuracy on an average with an error of 2.14 meter when all APs are considered. The selected dataset with minimal set of stable APs are found to give 90.13% average localization accuracy with an average error of 1.3 meter.
Year
Venue
Field
2018
ICDCN Workshops
Feature selection,Computer science,Location-based service,Real-time computing,Fingerprint,Supervised learning,Radio frequency,Metre (music),Distributed computing
DocType
Citations 
PageRank 
Conference
0
0.34
References 
Authors
9
2
Name
Order
Citations
PageRank
Priya Roy1122.54
Chandreyee Chowdhury23012.18