Affiliations: [a] Graduate School of Medicine and Engineering,
University of Yamanashi, Kofu, Japan | [b] Yokogawa Electric Corporation, Tokyo, Japan | [c] Graduate School of Information Sciences, Tohoku
University, Sendai, Japan | [d] Graduate School of Engineering, Tohoku University,
Sendai, Japan
Correspondence:
[*]
Corresponding author. Graduate School of Medicine and
Engineering, University of Yamanashi, 4-3-11 Takeda, Kofu, Yamanashi, Japan.
Tel./Fax: +81 55 220 8575; E-mail: yohtaki@rm.kanagawa-it.ac.jp
Abstract: This paper presents a method for indoor personal positioning
especially addressing on a simple gait measurement using a body-mounted
inertial sensor, a walk path estimation algorithm, and a trajectory
identification technique. Considering utility of on-site customer traffic
tracking, the method was designed to be autonomous needless of neither external
measures nor cumbersome installations to the environment. The inertial sensor
combining three-dimensional accelerometer, angular rate sensors, and
geomagnetic sensors was used to assess walking dynamics and to obtain changes
of heading directions. To avoid an inherent problem of the walk trajectory
estimation by conventional dead-reckoning algorithm, an advanced probabilistic
map matching method which employs a Particle filter was developed. The
experiment with healthy male adults was evaluated to confirm the utility of the
method in a condition of small retail store. The result showed that the
proposed method demonstrated sufficiently feasible tacking which delineated
personal trajectories and positioning while shopping. The method showed
possible availability with the portable instrument and improved estimation
accuracy by the advanced probabilistic map matching which identifies proper
routings on the basis of the maximum likelihood. The method will be useful to
enhance ambulatory personal tracking technique in daily indoor environment.
Keywords: Human walking, ambulatory measurement, inertial sensor, geomagnetic sensor, indoor positioning, location estimation, dead reckoning, map matching, particle filter
