2016 International Conference on Indoor Positioning and Indoor Navigation (IPIN), 4-7 October 2016, Alcalá de Henares, Spain
978-1-5090-2425-4/16/$31.00 ©2016 IEEE
Smallest Enclosing Circle-based Fingerprint
Clustering and Modified-WKNN Matching
Algorithm for Indoor Positioning
Wen Liu, Xiao Fu
, Zhongliang Deng, Lianming Xu, Jichao Jiao
School of Electronic Engineering
Beijing University of Posts and Telecommunications
Beijing, China
liuwen@bupt.edu.cn;
Xiaofu@bupt.edu.cn
;
dengzhl@bupt.edu.cn;
xulianming@rtmap.com;
jiaojichao@bupt.edu.cn
Abstract—Methods to cluster fingerprints based on Smallest-
Enclosing-Circle (SEC) and to modify Weighted-K-Nearest-
Neighbor (WKNN) matching algorithm for indoor fingerprint
positioning system are proposed. Based on the approach to
computing the smallest k-enclosing circle, the method proposed
clusters fingerprints in database by introducing reference points’
coordinates, instead of their received signal strength (RSS). This
approach performs higher accuracy of positioning areas
compared to conventional clustering algorithms, which are based
on RSS. Meanwhile, this paper analyses the transmission
characteristics of wireless signals in dense cluttered
environments, and derives a novel path-loss-model-based weight
computational method for WKNN matching algorithm. A
modified-WKNN (M-WKNN) matching algorithm for indoor
fingerprint positioning system is proposed and experiments are
implemented in China National Grand Theatre. Results show
that the location area accuracy using the proposed clustering
algorithm is improved by 30% compared to that using K-means
algorithm, and the positioning accuracy of M-WKNN is 11.9%
and 29.1% higher than that of WKNN and KNN, respectively.
Keywords—smallest enclosing circle; clustering algorithm;
matching principle; fingerprint positioning
I.
I
NTRODUCTION
With the rapid development of mobile communication
techniques, the demand for accurate localization in indoor
environments, such as department stores, office buildings and
transportation facilities, has increased dramatically [1, 2]. Since
the coverage of the satellite signal, GPS for example, is limited
in indoor environments, various technologies, including Wi-Fi
[3], Bluetooth [4], RFID [5], and Ultra-Wideband (UWB) [6]
have been proposed to provide better performance in indoor
localization. Fingerprint-based positioning technique has
aroused researchers’ interests because of its high mobility, low
networking cost and high compatibility in dense cluttered
indoor environment [7].
Large-scale indoor space, such as theatres, shopping malls
and exhibition halls, has complex environment in general,
which brings diversities to the transmission of wireless signals.
In other words, wireless signals’ transmission is not based on
free-space transmission model since there are dense multipath
and scatters in these indoor environment [8]. For example, two
reference nodes which are close to each other in physical-space
may have similar received signal strength (RSS) in signal-
space. Conventional clustering algorithms, take K-means [9]
for example, cluster reference points in fingerprint database
based on their received signal strengths, which cannot satisfy
the needs of high positioning area accuracy in such conditions.
This paper proposed a coordinates-based clustering method for
indoor fingerprint positioning system, using which can improve
localization area accuracy efficiently.
Matching algorithms for indoor fingerprint positioning
system are applied in the on-line stage. Traditional matching
algorithms, like K-Nearest Neighbor [10] (KNN) and Weighted
K-Nearest Neighbor [11] (WKNN), match the RSS between
experimental points and reference points in the database to
estimate the positions of the former. However, the matching
principles both in KNN and WKNN are not considerate about
the relation between RSS and real distance, which may cause
deviations for positioning system. To mitigate the deviations
above, a weight computational method for Modified-WKNN
(M-WKNN) is derived in this paper, based on path-loss-model.
II. E
RROR
A
NALYSIS OF
F
INGERPRINT
B
ASED
P
OSITIONING
S
YSTEM
The error in fingerprint positioning system mainly consists
of two parts, clustering error and matching error, which are
given by Equation (1).
cluster matching
Δ=Δ +Δ
(1)
Where ∆ is system’s total error, ∆
cluster
and ∆
matching
are the
errors included in clustering algorithms and matching
algorithms respectively.