Flying Trot Control Method for Quadruped Robot Based on Trajectory Planning

2025-05-06 0 0 6.31MB 35 页 10玖币
侵权投诉
Flying Trot Control Method for Quadruped Robot
Based on Trajectory Planning
Hongge Wanga, Hui Chai a,, Bin Chenb, Aizhen Xiea,b, Rui Songa, Bo Suc
aRobotics Research Center, School of Control Science and Engineering,Shandong
University, Shandong Province, China
bShandong Youbaote Intelligent Robot Co., Ltd, Jinan, Shandong Province, China
cChina North Vehicle Research Institute, Beijing, China
Abstract
An intuitive control method for the flying trot, which combines offline trajec-
tory planning with real-time balance control, is presented. The motion features
of running animals in the vertical direction were analysed using the spring-
load-inverted-pendulum (SLIP) model, and the foot trajectory of the robot was
planned, so the robot could run similar to an animal capable of vertical flight,
according to the given height and speed of the trunk. To improve the robustness
of running, a posture control method based on a foot acceleration adjustment
is proposed. A novel kinematic based CoM observation method and CoM regu-
lation method is present to enhance the stability of locomotion. To reduce the
impact force when the robot interacts with the environment, the virtual model
control method is used in the control of the foot trajectory to achieve active
compliance. By selecting the proper parameters for the virtual model, the os-
cillation motion of the virtual model and the planning motion of the support
foot are synchronized to avoid the large disturbance caused by the oscillation
motion of the virtual model in relation to the robot motion. The simulation and
experiment using the quadruped robot Billy are reported. In the experiment,
the maximum speed of the robot could reach 4.73 times the body length per
Corresponding author
Email addresses: wanghg1992@outlook.com (Hongge Wang ), chaimax@sdu.edu.cn (Hui
Chai ), 15820000319@139.com (Bin Chen ), xieaizhensdu@163.com (Aizhen Xie ),
rsong@sdu.edu.cn (Rui Song ), bosu@noveri.com.cn (Bo Su )
Preprint submitted to Journal of Mechatronics October 25, 2022
arXiv:2210.13446v1 [cs.RO] 24 Oct 2022
second, which verified the feasibility of the control method.
Keywords: Quadruped Robot, Flying Trot, Trajectory Planning, Posture
Control, Kinematic Based CoM Observation, Active Compliance Control
1. Introduction
The legged robot is more adaptable to the non-structural environment than
the wheeled robot and the tracked robot. The quadruped robot has better sta-
bility than the biped robot and has a simpler structural complexity than the
hexapod robot, so the quadruped robots have attracted increasingly more at-
tention from scholars.
Improving the speed of movement is a trend in quadruped robots. The
common gaits of quadruped robots include crawl, trot, pace, bound, and gallop.
Due to the good symmetry, small posture fluctuations, and the ability to balance
the movement speed and energy consumption, the trot gait is widely used in
the movement of the quadruped robot.
The flying trot gait is a trot gait with flying phases that can achieve higher
moving speeds. The quadruped robots that run with a flying trot gait are
Quadruped, designed by the Massachusetts Institute of Technology [1, 2], Kolt
[3, 4, 5] of Stanford University and Ohio State University, HyQ [6] of the Italian
Institute of Technology, StarlETH [7, 8] of the Zurich Federal Institute of Tech-
nology, Cheetah cub [9] of Lausanne Federal Institute of Technology, Boston
Dynamics’ BigDog [10, 11], LS3 [12] and Spot [13]; in addition, Billy [14] is
the quadruped robot used in the experiments of this paper, as shown in Fig.
1a, which is cooperatively developed by Shandong University and Shandong
Youbaote Intelligent Robot Co., Ltd. The control methods of the flying trot
gait that is used are summarized as follows.
The researchers of Quadruped put forward the control method, which di-
vided the control of the robot into three parts: trunk height control, posture
control and speed control, and realized a stable trot movement. Controlling the
2
(a) Billy quadruped robot (b) Simulation model
Figure 1: Billy, the quadruped robot used in the experiments, and its simulation model in
Webots
energy of the robot in a vertical direction realized the flying of the Quadruped
robot in the trot movement [1, 15, 16]. However, because the control method
of vertical flying is closely combined with the pneumatic control method of the
legs, it is not universal.
By controlling the touch position of the swing foot to control the robot’s
forward velocity, lateral velocity, heading angular velocity, in addition to con-
trolling the energy injected to the support leg spring to control the trunk height
and controlling the joints’ torque of the support leg to control the pitch angle
and roll angle, Kolt can move stably with a trot gait [8]. When the descent
speed of the support foot reaches a certain value, Kolt uplifts the support foot
to achieve flying in the trot gait [9]. However, the researchers did not detail the
specific process of the descent control and uplift control of the support foot, nor
did they establish the specific model of the flying control.
The motion control of the quadruped robot StarlETH is divided into trunk
control, swing foot control and support foot control. The swing foot control is
mainly used to select the proper touch position to control the robot speed and
make the robot recover from an external disturbance. The trunk control mainly
calculates a desired virtual six-dimensional force acting on the trunk from the
deviation between the control target and the actual situation of trunk and then
3
calculates a group of optimal forces acting on the support feet through an opti-
mization method. The support foot control aims to calculate the corresponding
joint torque from the desired optimal foot forces of the support legs and execute
the torque. The flying phase is achieved by controlling the height of the trunk
in the trot gait [8]. However, the researchers of StarlETH did not details the
process of the height control of the trunk to achieve the flying phase.
The quadruped robot Cheetah cub uses open-loop control. Its researchers
control the movement of the quadruped robot with the method of a central
pattern generator (CPG) based on the joint workspace. Then, through a large
number of experiments, the parameters of CPG are optimized to improve the
forward speed of the quadruped robot [9]. However, the optimal parameters
of different quadruped robots are different, and the process of parameter op-
timization is complex. Moreover, since the quadruped robot uses the CPG
method based on joint space, this method may not be applicable once the joint
configuration of the robot is changed.
The flying trot of the quadruped robot HyQ is based on active compliant
control. The desired trajectory of the foot is planned by the CPG method based
on the foot workspace [17], and the desired joint angle is obtained by inverse
kinematics. PD control is then used in joint space to make the actual joint angle
follow the desired joint angle and achieve active compliance. They stabilized
the flying trot by choosing an appropriate step frequency and duty factor [18]
to make the motion of the quadruped robot synchronize with the oscillation
of the virtual spring-mass system in the support phase [6]. However, the HyQ
researchers did not provide the method for selecting the step frequency and duty
factor.
The flying trot control methods of BigDog, LS3, and Spot have not been
publicly disclosed. The current information of a quadruped robot that can run
with a flying trot gait is shown in table 1. Cheetah cub has a ratio of speed
to body length greater than Billy, but it has no the rolling rotary joint, so it
cannot move flexibly in all directions, and it is lighter because its power supply
is external.
4
Table 1: Flying trot speed of quadruped robots
Year Robot Mass (kg) BL (m) Speed (BL/s)
1990 Quadruped 38 0.78 2.82
2004 Kolt 80 1.75 0.63
2010 HyQ 70 1.3 1.3
2013 StarlETH 23 0.5 *
2013 Cheetah cub 1.1 0.205 6.9
BigDog 109 1.1 1.82
2004-2015 LS3 * * *
Spot 72 * *
2016 Billy 1.9 0.167 4.73
This paper takes quadruped as the research object, analyses the movement
process of a quadruped in the running of the flying trot gait with the SLIP
model, and puts forward the gait control method of the flying trot based on
the position planning of the foot workspace, including the control of the robot’s
eject, landing buffer and forward speed. To increase the robustness of movement,
a new posture control method based on position was proposed, and the method
of virtual model control was used to carry out active compliance control to
reduce the impact force when the robot interacts with the environment.
The advantages of the method proposed in this paper are as follows. (1) The
implementation method of the flying trot gait based on position planning was
proposed, and the touchdown time was predicted in planning to make buffer
actions in advance and reduce the impact of landing. (2) A novel position-
based posture control method was proposed, which can be used to enhance the
balance of the robot. (3) To further reduce the impact of landing in running,
the method of the leg virtual model was used for active compliance control, and
a reasonable selection method of kp and kd for virtual spring and damping was
proposed for the running gait.
The simulations in Webots and the experiments of Billy prove that the
5
摘要:

FlyingTrotControlMethodforQuadrupedRobotBasedonTrajectoryPlanningHonggeWanga,HuiChaia,,BinChenb,AizhenXiea,b,RuiSonga,BoSucaRoboticsResearchCenter,SchoolofControlScienceandEngineering,ShandongUniversity,ShandongProvince,ChinabShandongYoubaoteIntelligentRobotCo.,Ltd,Jinan,ShandongProvince,ChinacChin...

展开>> 收起<<
Flying Trot Control Method for Quadruped Robot Based on Trajectory Planning.pdf

共35页,预览5页

还剩页未读, 继续阅读

声明:本站为文档C2C交易模式,即用户上传的文档直接被用户下载,本站只是中间服务平台,本站所有文档下载所得的收益归上传人(含作者)所有。玖贝云文库仅提供信息存储空间,仅对用户上传内容的表现方式做保护处理,对上载内容本身不做任何修改或编辑。若文档所含内容侵犯了您的版权或隐私,请立即通知玖贝云文库,我们立即给予删除!
分类:图书资源 价格:10玖币 属性:35 页 大小:6.31MB 格式:PDF 时间:2025-05-06

开通VIP享超值会员特权

  • 多端同步记录
  • 高速下载文档
  • 免费文档工具
  • 分享文档赚钱
  • 每日登录抽奖
  • 优质衍生服务
/ 35
客服
关注