Abstract— How to evaluate the walking stability of robots and
develop stable and effective walking control algorithms is a key
research hotspot in the development of humanoid biped robots.
At present, the main stability evaluation basis is zero moment
point stability criterion. However, this method is poor in stability,
flexibility and efficiency, which severely limits the walking
ability of humanoid biped robot. This article is based on the
human walking motion capture data, combining with the
anatomical model, the kinematics forward solution algorithm
which is used to calculate the trajectory of the human joint
establish the kinematics model of the human walking process.
This model analyzes the angular momentum balance and moment
balance by introducing ground reference points and combining
Euler's second theorem. Then on the base of these, we selected
the forward distance between zero moment point and central
moment axis, the vector plane centroid angular momentum and
the acceleration of the center of mass in the forward direction as
parameters to establish the human gait stability evaluation rules .
Finally, The simulation results show that this stability criterion
achieves an accurate discrimination of the walking gait stability
of the human body and has high accuracy in determining the gait
stability.
I. INTRODUCTION
At present, with the development of robot research, more
and more humanoid robots are beginning to enter the market,
such as the occasional consulting robots in public places, the
bank hall, the trading hall and other public places. Some
children are available for communication, have some teaching
and auxiliary ability to accompany the robot; some restaurants
have been put into use the service robot and so on. However,
robots that actually have the ability to walk biped and walk like
humans have never been on the market.The Professor Kato
Ichiro of the Waseda University in Japan once proposed that
"one of the biggest functions a robot should have is a walking
function"[1]. Nowadays, bipedal humanoid robots are
becoming more and more important. Inspired by the disaster
relief difficulties after the Fukushima nuclear power plant
accident, the United States has hosted the DARPA (Defense
Advanced Research Projects Agency) disaster-relief robot
challenge[2,3]. In the DARPA competition, the disaster relief
robots exhibited by the teams saw the possibility of bipedal
humanoid robots replacing humans in hazardous environments;
The series of robots of Atlas developed by the world's leading
robotics company, Boston Dynamics, USA, has now achieved
Xiaoguang Wu, Institute of Electric Engineering, Yanshan University,
Heibei,China. E-mail: wuxiaoguang@ysu.edu.cn. supported by the National
Natural Science Foundation of China (NSFC) under Grants
61503325/F2014203246, Postdoctoral Science Foundation of China under
Grants 2015M581316 and the self-managed project of YANSHAN
University under Grants 13LGB009.
difficult motions such as anti-disturbance, handling and
backflip. However, at this stage, the problem of walking
stability is still a huge problem between the biped robot and the
market.
In order to achieve stable biped gait for active robots, the
researchers put forward various stability criteria with different
adaption. Such as, the stability criterion of the Center of
Pressure ,the stability criterion of the Zero Moment Point[4],
according to the human body's centroid angular momentum
analysis the Foot Rotation Indicator stability criterion was
introduced by Goswami et al[5,6]. But with the development
of robotics,the problem of poor robustness and poor
compliance of ZMP stability criterion is becoming more and
more serious, it is urgent to find new breakthroughs. The
existing passive biped robot control algorithms are also based
on simpler robot models,which are greatly influenced by the
robot body[7,8].
Bipedal walking is a basic activity of the human body[9],
after a long period of evolution, the human body has become
the most perfect bipedal motion system in nature. An in-depth
study of the principle of stable walking of the human body can
help us to understand the mechanism of dynamic walking of
the bipedal feet and to develop humanoid bipedal robots.
Based on the above reasons, this article intends to study the
stability criteria of the bipedal system, proposing a gait
stability criterion based on cognition of human walking
balance mechanism, which will lay a foundation for further
development of the imbalance detection and control algorithm
for humanoid biped robots and provide a new direction for gait
recognition of walking disabled patients.
II. HUMAN MODEL DIVISION BASED ON MOTION CAPTURE
DATA
In order to analyze the walking stability of the human body,
first, by wearing inertial sensor records the location
information or use camera instrument to record the location of
the optical reflection point motion capture experiments with
real time kinematic data in the process of walking. Based on
the idea of bionics, the human body's physiological
organization structure is used to establish a simplified model
of the human body and divide the model into segments.
According to human morphology[10,11,12],the kinematics
model is divided into five kinematic chains: left foot, right foot,
spine, left arm and right arm. Each kinematic chain is
composed of a set of rigid body connecting rods that connect
the joints. The common starting point of all kinematic chains is
the root node of the hip joint center. Based on the human
motion model, a method proposed by Denavt and Harttenberg
Analysis of Gait Stability Criterion Based on Cognition
of Human Walking Balance Mechanism*
Xiaoguang Wu, Lei Yang, Shaowei Liu, Tingjin Wang, Yanhui Li and Lei Wei
Proceedings of the 1st WRC Symposium
on Advanced Robotics and Automation 2018
Beijing, China, August 16, 2018
978-1-5386-7674-5/18/$31.00 ©2018 IEEE 34