Brian
G Mmari
SURF Active Aerodynamic Tail Enhances Agile Locomotion of Legged Robots in Challenging Natural Environments Innovative Technology / Entrepreneurship / Design
Abstract profile. Full document pending author claim.
Authors:
Brian G Mmari
Date Created:
Not specified
Course Title:
Professor:
Not specified
About Paper:
Terrestrial animals have evolved over millennia to effortlessly locomote across complex terrains, such as grasslands and forests. They exhibit a variety of gait patterns such as hopping, running, and even mid-air reorientation. A common feature among these animals is their tail which helps them in dynamic stabilization through the various gait patterns. While tails are being developed to improve the performance and adaptability of legged quadruped robots, they are modeled as passive mass which can only provide limited corrective torques to help in dynamic stabilization. In this work, we develop an Active Aerodynamic Tail (AA-Tail), with two propellers integrated at the tail end to enhance agile locomotion performances of legged robots. The proposed design weighs around 150g and enables highly precise and fast reaction capabilities, enhancing the robot's dynamic stabilization. For validation, we develop and deploy the tail on a 500g quadruped robot dog, Mini Pupper 2. The simulations and experiments demonstrate that our tail effectively enhanced the legged robot's agility across various motion patterns simultaneously, including running, turning, leaping, slope running, standing up, sidestepping, and forward and backward locomotion. This approach proves the effectiveness of the tail as an active stabilization component essential for robots to mimic natural gait patterns in complex terrains. Keywords: Robotics; Legged Robots; Reinforcement Learning
Source:
Purdue University / 2024
Topics:
No topics listed
Co-authors:
Brian G Mmari