1
Wearable Indoor Pedestrian Navigation Based on MIMU and Hy-
pothesis Testing
*
Xiao-fei Ma
1
, Zhong Su
†‡2
, Xu Zhao
1
, Fuchao Liu
1
, Chao Li
1
(
1
School of Automation, Beijing Institute of Technology, Beijing, 100084, China)
(
2
Beijing Key Laboratory of High Dynamic Navigation Technology, Beijing Information Science & Technological University, Beijing, 100101,
China)
Email-Address: jzys8000@163.com
Abstract: Indoor pedestrian navigation (IPN) has attracted more and more attention for the reason that it can be widely used in
indoor environments without GPS, such as fire and rescue in building, underground parking, etc. Pedestrian Dead Reckoning
(PDR) based on inertial measurement unit just to meet the demand in this. This paper designs and implements a miniature wearable
indoor pedestrian navigation system to estimate the position and attitude of a person while walking indoor. In order to reduce the
accumulated error due to long-term drift of inertial devices, a zero-velocity detector based on hypothesis testing is introduced for
instantaneous velocity and angular velocity correction. A Kalman filter combining INS information, magnetic information and
zero transient correction information is designed to estimate system errors and correct them. Finally, performance testing and
evaluation are conducted to the IPN; the results show that for leveled ground, a position accuracy is about 2% of the travelled
distance.
Key words: Wearable Indoor Pedestrian Navigation, MIMU, ZUPT, EKF, hypothesis testing
doi: Document code: A CLC number:
1 Introduction
Indoor pedestrian navigation (IPN) system can
track and locate indoor pedestrian, which can be
widely used in location inside complex building,
firefighters rescue at the fire scene, robot path plan-
ning, etc. The main difficulty is that a high position-
ing accuracy GPS is difficult to use like the outdoor
environment. There are two main kinds of IPN pattern
which are preset-node pattern and autonomous pat-
tern.
Preset-node pattern means some positioning
nodes (like beacon node, base station) need to be
arranged before location. Representative technologies
include ultra-wideband wireless location (UWB),
Bluetooth positioning (Bluetooth), WLAN position-
ing (WLAN), radio frequency identification posi-
tioning (RFID), ZigBee positioning, ultrasonic posi-
tioning, positioning infrared near-field electromag-
netic positioning (NFER), etc. This method has a high
precision, but requires high-density preset nodes in
the navigation area. Not only high cost, it is unable to
locate overall environment due to some blind zone in
real time, and positioning may failure when some
node can’t effectively service.
Autonomous navigation is a method without
pre-node. It is mainly use MEMS inertial measure-
ment unit (MIMU, mainly includes three-axis gyro-
scopes, three-axis accelerometer, even magnetic and
pressure sensors) mounted on the pedestrian body
(mostly feet) to obtain information on pedestrian
movement, geomagnetism and others, then estimate
attitude and location by dead reckoning or strapdown
resolving. This method is well-adapted without preset
beacon, but its accuracy related to its running time
and distance. At present, research on this problem is
mainly on how to increase correction information in
existing algorithm framework, and improve infor-
mation fusion filter.
This paper designs and implements a miniature
wearable indoor pedestrian navigation system, the
‡
Corresponding author
*
Project supported by the National Natural Science Foundation of
China (No. 61471046)