Whole-body kinematic and dynamic modeling for quadruped robot under different gaits and mechanism topologies

Dynamic locomotion plays a crucial role for legged robots to fulfill tasks in unstructured environments. This paper proposes whole-body kinematic and dynamic modeling method s based on screw theory for a quadruped robot using different gaits and mechanism topologies. Unlike simplified models such as...

Full description

Saved in:
Bibliographic Details
Published in:PeerJ. Computer science 2021-12, Vol.7, p.e821-e821, Article e821
Main Authors: Yan, Wei, Pan, Yang, Che, Junjie, Yu, Jiexian, Han, Zhuchen
Format: Article
Language:English
Subjects:
Analysis
Artificial Intelligence
Centroids
Control systems
Design
Dynamic models
Embedded Computing
End effectors
Gait
Inertia
Inertial platforms
Kinematics
Locomotion
Mechanism topologies
Quadruped robot
Real-Time and Embedded Systems
Robotics
Robots
Screw theory
Sensors
Theory and Formal Methods
Topology
Velocity
Walking
Whole body dynamic
Citations: Items that this one cites
Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Dynamic locomotion plays a crucial role for legged robots to fulfill tasks in unstructured environments. This paper proposes whole-body kinematic and dynamic modeling method s based on screw theory for a quadruped robot using different gaits and mechanism topologies. Unlike simplified models such as centroid or inverse pendulum models, the methods proposed here can handle 10-dimensional mass and inertia for each part. The only simplification is that foot contact models are treated as spherical joints. Models of three different mechanism topologies are formulated: (1) Standing phase: a system consisting of one end-effector, the body, and four limbs, the legs; (2) Walking phase: a system consisting of one or two lifting legs (depending on the chosen gait), two or three supporting legs; (3) Floating phase: a system in which all legs detach from the ground. Control strategies based on our models are also introduced, which includes walk and trot gait plans. In our control system, two additional types of information are provided: (1) contacting forces are given by force sensors installed under feet; (2) body poses are determined by an inertial measurement unit (IMU). Combined with the sensor data and calibrated mass, inertia, and friction, the joint torque can be estimated accurately in simulation and experiment. Our prototype, the "XiLing" robot, is built to verify the methods proposed in this paper, and the results show that the models can be solved quickly and leads to steady locomotions.
ISSN:2376-5992
2376-5992
DOI:10.7717/peerj-cs.821