Auton Robot
DOI 10.1007/s10514-015-9433-4
Biologically inspired gait transition control for a quadruped
walking robot
Ig Mo Koo · Tran Duc Trong · Yoon Haeng Lee ·
Hyungpil Moon · Jachoon Koo · Sangdoek Park ·
Hyouk Ryeol Choi
Received: 20 February 2013 / Accepted: 6 January 2015
© Springer Science+Business Media New York 2015
Abstract The gait transition of a quadruped walking robot
is the switching of gait with non-periodic gait sequences
between the periodic ones such as from walk to trot or trot to
walk etc. It is very much important because the robot should
change its gait depending upon the moving speed to enhance
the efficiency of locomotion. In this paper, we present a quasi-
static gait transition control method for a quadruped walk-
ing robot. It is based on the observation on the locomotion
behaviors of quadruped animals, which show a sudden and
discrete changes of gait patterns depending on the speed. The
method predefines gait transition patterns, and gait sequences
are determined according to the current and desired leg pos-
Electronic supplementary material The online version of this
article (doi:10.1007/s10514-015-9433-4) contains supplementary
material, which is available to authorized users.
I. M. Koo
University of Texas, Houston, TX, USA
e-mail: kooigmo@gmail.com
T. D. Trong · Y. H. Lee · H. Moon · J. Koo · H. R. Choi (B )
School of Mechanical Engineering, Sungkyunkwan University,
Chonch’on-dong, Jangan-gu, Suwon, Kyonggi-do, Korea
e-mail: hrchoi@me.skku.ac.kr
T. D. Trong
e-mail: jamestran@me.skku.ac.kr
Y. H. Lee
e-mail: yhlee@me.skku.ac.kr
H. Moon
e-mail: hyungpil@me.skku.ac.kr
J. Koo
e-mail: jckoo@me.skku.ac.kr
S. Park
Division of Applied Robot Technology, Korea Institute of Industrial
Technology, Ansan, Korea
e-mail: sdpark88@gmail.com
tures. It can be useful because the applicable to any type
of walking controller. In this study, we implement the pro-
posed method on a self-contained quadruped walking robot,
called Artificial Digitigrade for Natural Environment Ver-
sion III (AiDIN-III), and its effectiveness is experimentally
validated.
Keywords Quadruped walking robot · Gait transition ·
Leg control · Sequence
1 Introduction
The study of quadruped walking robots is a kind of clas-
sic issues in robotics and it has been performed by many
researchers for long time (Raibert et al. 2008; Jinpoong Jin-
poong; Rebula et al. 2007; Kimura and Fukuoka 2004; Koo
et al. 2007, 2009). Recently, the importance of quadruped
walking robots is emphasized again in accordance with the
increasing necessity of robots with excellent mobility and
adaptability in natural environments.
Along with the mechanical structure of a quadruped walk-
ing robot, its walking control is one of the most important
issues in order to realize successful locomotion. Up to now
various control methods have been reported, for example,
trajectory-based control (Yoneda and Hirose 1995; Hunang
et al. 2001; Choi et al. 2004), heuristic control (Chew et al.
1998; Pratt et al. 2001), and oscillatory reflex control (Kimura
and Fukuoka 2004; Koo et al. 2009; Tran et al. 2009; Berns
et al. 1999; Takeuchi 1999; Allen et al. 2003) etc. How-
ever, most of them just focus on single gait pattern such as
walk or trot etc. It works well if the environmental condi-
tions do not greatly change and the required speed of loco-
motion is constant. On the contrary, it is not appropriate in
natural environments such as irregular, and rocky terrains.
123