A Hybrid Step Model and New Azimuth Estimation
Method for Pedestrian Dead Reckoning
Yaping Zhu, Rui Zhang, Weiwei Xia, Ziyan Jia and Lianfeng Shen
National Mobile Communications Research Laboratory
Southeast University
Nanjing, Jiangsu 210096, China
Email:{xyzzyp, zhangrui09, wwxia, lfshen}@seu.edu.cn, jiaziyan@jsut.edu.cn
Abstract — This paper proposes a hybrid step length model
and a new azimuth estimation method based on the Pedestrian
Dead Reckoning for personal indoor positioning. The hybrid
step length model combines the information of the walking
frequency and the variance of accelerations with the
amplitudes of the acceleration signal. By weighting the azimuth
estimated through the gyroscopes and the magnetometers
respectively according to the angular rates, the accuracy of the
azimuth estimation is improved.
Keywords—hybrid step length model, azimuth estimation,
Pedestrian Dead Reckoning.
I. INTRODUCTION
In recent years, the rapid development of data services
has contributed to people’s growing demand for
Location-Based Services (LBS) [1]. The Global Navigation
Satellite System (GNSS) [2] which is widely used now can
almost cover globally, however, since the ability of the
signal to penetrate is weak, the positioning accuracy of the
receiver is poor, sometimes it is unable to locate in indoor
environments. And therefore how to provide a more precise
positioning method in indoor environments becomes one of
comparing urgent needs.
During the last decade several methodologies have
been proposed for accurate pedestrian position estimation
based on inertial sensors [3], [4]. These methodologies,
called Pedestrian Dead Reckoning (PDR) [5], compute a
person’s position through step detection, step length
calculation and orientation estimation, which is often used
in the Inertial Navigation System (INS). The Inertial
measurement Units (IMU) is usually composed of a set of
wearable sensors which include sensors such as
accelerometers, gyroscopes, electronic compasses,
barometers, temperature sensors, etc. A tri-axial
accelerometer is a fundamental aiding sensor for step and
attitude estimation. A novel method to overcome
nonidealities such as scale factors, cross coupling, bias, and
This work is supported in part by the research project of electric power
optical fiber sensing, information sensing and optical wireless unified
communication technology, the National Natural Science Foundation of
China (No.61201175, 61201248), the Important National Science &
Technology Specific Projects (No.2012ZX03004005-003), and Changzhou
Planning Project of Science and Technology (No. CJ20120030).
other higher-order nonlinearities which affects the output of
accelerometer is proposed in [6]. Seong-hoon Peter Won
and Farid Golnaraghi propose a new tri-axial accelerometer
calibration method using a mathematical model of six
calibration parameters in [7].Gyroscopes providing angular
rate are used for orientation which may cause heading drift,
a method integrating Zero Angular Rate Updates (ZARU) ,
Heuristic Heading Reduction (HDR) and Electronic
Compass into a Kalman-based INS-EKF-ZUPT (IEZ)
platform can be seen in [8], which is aimed at reducing the
heading drift. Literature [9] develops a novel Quasi-Static
magnetic Field (QSF) based attitude and angular rate error
estimation techniques to effectively use magnetic
measurements in highly perturbed environments.
In this paper, we propose a hybrid step length model
and a new azimuth calculation method for PDR based on
inertial sensors, which are composed of a triad of orthogonal
accelerometers, a triad of orthogonal gyroscopes and a triad
of orthogonal magnetometers. The experimentation results
show that our improved PDR algorithm which adopted a
hybrid step length model and new azimuth estimation
method outperforms the raw Dead Reckoning (DR)
algorithm with 0.3249m in maximum positioning error and
0.1504m in mean.
II. HARDWARE
DESCRIPTION
The equipment used in this paper consists of the
sensing module, acquisition module and a PC. The system
block diagram is shown as Figure 1. The sensing module is
the 9DOF Razor IMU which incorporates three sensors- an
ITG-3200 (MEMS triple-axis gyroscope), ADXL345
(triple-axis accelerometer), and HMC5883L (triple-axis
magnetometer), to give nine degrees of inertial
measurement.
The inertial measurement includes the acceleration data,
angular rate and magnetic field strength in three dimensions.
In addition, the outputs of all sensors are processed by an
on-board ATmega328 and output over a serial interface.
This enables the 9DOF Razor to be used as a very powerful
control mechanism for UAVs, autonomous vehicles and
image stabilization systems. The inertial data is received by
the Arduino Mega Board via the serial TX and RX pins with
a speed of 57600bps.
2014 Sixth International Conference on Wireless Communications and Signal Processing (WCSP)
978-1-4799-7339-2/14/$31.00 ©2014 IEEE