In the growing field of legged robotics, the integration of locomotion and manipulation—loco-manipulation—poses a significant challenge, addressed by leading organizations such as Boston Dynamics, Tesla, and Unitree. This work presents an open-source framework that facilitates planning through contact, enhancing robots’ capabilities in logistics, construction, and even planetary exploration. Our approach leverages a versatile trajectory optimization method and real-time perception to generate dynamic motions, coordinated through whole-body control. We validate our framework’s effectiveness on the Unitree Go1 equipped with a custom arm, demonstrating its application across various gaits and jumping maneuvers, as well as in simulated environments on humanoid robots like HURON and Atlas. The results showcase the framework’s potential to enable complex athletic tasks and manipulation while navigating diverse terrains.

• This report represents the work of one or more WPI undergraduate students submitted to the faculty as evidence of completion of a degree requirement. WPI routinely publishes these reports on its website without editorial or peer review.

Creator

• Wang, Lehong
• Jaitly, Akshay
• Yuan, Yifu
• Zou, Tao
• Chandler, Ethan
• Xu, Puen

Subject

• Project-Based Learning
• Industry
• Infrastructure
• Computing
• Security
• Innovation

Publisher

• Worcester Polytechnic Institute

Identifier

• E-project-042524-211648
• 121844

Palabra Clave

• Optimization
• Perception
• Open source
• Legged robots
• Control
• Hardware

Advisor

• Xiao, Jing
• Rosendo, Andre
• Michalson, William R.
• Agheli Hajiabadi, Mohammad Mahdi

Year

• 2024

Sponsor

• Unitree Robotics
• MathAltitude School of Mathematics

UN Sustainable Development Goals

• Not Specified

Date created

• 2024-04-25

Resource type

• Major Qualifying Project

Major

• Robotics Engineering

Source

• E-project-042524-211648

Rights statement

• In Copyright

Última modificación

• 2024-05-21