Y. Cai et al. (Eds.): Mobilware 2010, LNICST 48, pp. 313–325, 2010.
© Institute for Computer Sciences, Social Informatics and Telecommunications Engineering 2010
Predicted and Corrected Location Estimation of Mobile
Nodes Based on the Combination of Kalman Filter and
the Bayesian Decision Theory
Muhammad Alam
1,2
, Mazliham Muhammad Suud
1
, Patrice Boursier
2
,
Shahrulniza Musa
1
, and Jawahir Che Mustapha Yusuf
1,2
1
Centre for Research and Postgraduate Studies (CRPGS) & UniKL MIIT
Jln Sultan Ismail, 50250, Kuala Lumpur
muhammad.unikl@gmail.com, mazliham@unikl.edu.my,
{shahrulniza,jawahir}@miit.unikl.edu.my
2
Laboratoire L3i, Université de La Rochelle, 17000 La Rochelle, France
patrice.boursier@univ-lr.fr
Abstract. The main objective of this research is to apply statistical location
estimation techniques in cellular networks in order to calculate the precise
location of the mobile node. Current research is focusing on the combination of
Kalman filter and the Bayesian decision theory based location estimation. In
this research basic four steps of Kalman filter are followed which are
Estimation, Filtering, Prediction and Fusion. Estimation is done by using
Receive Signal Strength (RSS), Available Signal Strength (ASS) and the Angle
of Arrival (AOA). Filtering is done by calculating the average location and
variation in values of location. Prediction is done by using the Bayesian
decision theory. Fusion is done by combining the variances calculated in
filtering step. Finally by combining the prediction and fusion results PCLEA
(Predicted and Corrected Location Estimation Algorithm) is established.
Timestamp is used for recursive step in kalman filter. The aim of this research
is to minimize the dependence on the satellite based location estimation and
increase its accuracy, efficiency and reliability.
Keywords: Kalman filter, Bayesian decision theory, location estimation.
1 Introduction
Location estimation of a mobile user is a very popular research area from past few
years. Due to the growth of cellular architecture the mobile users originating calls are
also increasing at the same time. It is estimated that more than 50% emergency calls
are originated by the mobile phones [1]. Techniques which are used for location
estimation are satellite based techniques, geometric techniques, statistical techniques
and the mapping techniques [2], [3]. All techniques have different accuracy level,
processing time, coverage and the cost. The location of the mobile node can be
estimated by the mobile node itself which is known as self positioning. Otherwise it
can be calculated by the server with the help of the reference points, which is known as
remote positioning or network centric positioning [4]. Two different approaches are
used by the researchers, the direct positioning approach and the two step-step