수동적으로 연결된 미지의 물체의 운동량 관리 통합을 통한 관절형 다리-팔 로봇의 균형유지 전신제어
Whole-Body Control for Balancing an Articulated Legged-Arm Robot with the Integration of Momentum Management and a Passively Connected Unknown Object
0
조회
0
다운로드
0
인용
0
참고문헌
키워드
초록
Maintaining balance is a critical challenge for legged mobile manipulators, especially when unknown objects are passively connected, as they can disrupt overall stability. To address this, we propose a whole-body control framework that incorporates momentum-based management to enhance balance in a torque-controlled quadrupedal manipulator. Our approach integrates momentum regulation directly into the control strategy, enabling the robot to account for the dynamic effects of connected objects without requiring separate control algorithms. This facilitates effective handling of dynamic interactions between the robot’s body and its components. Consequently, the robot’s ability to maintain balance during complex tasks—such as ensuring contact stability and tracking end-effector trajectories—is significantly improved. We demonstrate the effectiveness of this framework through simulations, where an articulated legged-arm robot maintains balance on an incline while tracking a desired trajectory with an object whose center of mass changes dynamically and is suspended from its arm. The proposed controller reduced position RMSE by 88% and Euler angle error by 77%, demonstrating its effectiveness.
Abstract
Maintaining balance is a critical challenge for legged mobile manipulators, especially when unknown objects are passively connected, as they can disrupt overall stability. To address this, we propose a whole-body control framework that incorporates momentum-based management to enhance balance in a torque-controlled quadrupedal manipulator. Our approach integrates momentum regulation directly into the control strategy, enabling the robot to account for the dynamic effects of connected objects without requiring separate control algorithms. This facilitates effective handling of dynamic interactions between the robot’s body and its components. Consequently, the robot’s ability to maintain balance during complex tasks—such as ensuring contact stability and tracking end-effector trajectories—is significantly improved. We demonstrate the effectiveness of this framework through simulations, where an articulated legged-arm robot maintains balance on an incline while tracking a desired trajectory with an object whose center of mass changes dynamically and is suspended from its arm. The proposed controller reduced position RMSE by 88% and Euler angle error by 77%, demonstrating its effectiveness.