Positioning with Bayesian coverage area estimates and location fingerprints
KOSKI, LAURA (2010)
KOSKI, LAURA
2010
Matematiikka - Mathematics
Informaatiotieteiden tiedekunta - Faculty of Information Sciences
This publication is copyrighted. You may download, display and print it for Your own personal use. Commercial use is prohibited.
Hyväksymispäivämäärä
2010-04-09
Julkaisun pysyvä osoite on
https://urn.fi/urn:nbn:fi:uta-1-20461
https://urn.fi/urn:nbn:fi:uta-1-20461
Tiivistelmä
A variety of commercial location based services have appeared during recent years. Location awareness is becoming more important also in environments where satellite-based positioning are not available, such as urban areas and indoors. In this work, a method to estimate the coverage area of a wireless communication node is presented. Also a method to use a database of such coverage area estimates for personal positioning is presented. Coverage area estimates are computed using location fingerprinting.
The coverage area is solved by forming the posterior distribution of the parameters using Bayes' rule. The coverage area of a communication node is modeled as an ellipse and is assumed to follow a multivariate normal linear model, which is presented as a general case. The coverage area estimate is derived using both noninformative and informative priors. Also a model which assumes a possibility of outliers and a Bayesian method for detecting outliers are presented.
A positioning method which uses the coverage area estimates is presented. The distribution of a position estimate is derived using Bayes' rule. The position estimate is weighted average of the centers of ellipses and the weights are determined by the shape parameters of ellipses.
Finally, the accuracy and consistency of a position estimate are studied using different coverage estimates.
Asiasanat:multivariate linear models, Bayesian analysis, location fingerprinting, coverage area estimation
The coverage area is solved by forming the posterior distribution of the parameters using Bayes' rule. The coverage area of a communication node is modeled as an ellipse and is assumed to follow a multivariate normal linear model, which is presented as a general case. The coverage area estimate is derived using both noninformative and informative priors. Also a model which assumes a possibility of outliers and a Bayesian method for detecting outliers are presented.
A positioning method which uses the coverage area estimates is presented. The distribution of a position estimate is derived using Bayes' rule. The position estimate is weighted average of the centers of ellipses and the weights are determined by the shape parameters of ellipses.
Finally, the accuracy and consistency of a position estimate are studied using different coverage estimates.
Asiasanat:multivariate linear models, Bayesian analysis, location fingerprinting, coverage area estimation