Biomedical Engineering Reference
In-Depth Information
Chapter 8
Sensing Human Walking: Algorithms
and Techniques for Extracting
and Modeling Locomotion
Franck Multon
Abstract
This chapter reports the most popular methods used to evaluate the main
properties of human walking. We will mainly focus on: global parameters (such as
step length, frequency, gait asymmetry and regularity), kinematic parameters (such
as joint angles depending on time), dynamic values (such as the ground reaction
force and the joint torques) and muscle activity (such as muscle tension). A large set
of sensors have been introduced in order to analyze human walking in biomechanics
and other connected domains such as robotics, human motion sciences, computer
animation… Among all these sensors, we will focus on: mono-point sensors (such
as accelerometers), multi-point sensors (such as flock of sensors, opto-electronic
systems and video analysis), and dynamic sensors (such as force plates or elec-
tromyographic sensors). For the most popular systems, we will describe the most
popular methods and algorithms used to compute the parameters described above.
All along the chapter we will explain how these algorithms could provide original
methods for helping people to design natural navigation in VR.
8.1 Introduction
Measurement of human motion has a long history since J.E. Marey and E. Muybridge
have proposed chronophotography to record animals inmotion [
20
]. Motionwas then
expressed as a sequence of poses which is still widely used to describe human motion
nowadays. Numerous sensors have been introduced to capture human motion and
especially human walking. In this chapter we propose a summary of the most studied
parameters in human walking and direct/indirect methods to obtain them.
Human walking is one of the most commonly used motion in everyday life and it
has been studied from a long time. However it is still a very active field of research
(
B
)
University Rennes2, Campus La Harpe, Av. Charles Tillon, CS24414, F35044 Rennes, France
e-mail: franck.multon@univ-rennes2.fr
F. Multon
