Khái niệm cốt lõi
Novel framework integrates MPC and pose optimization for effective loco-manipulation of heavy objects.
Tóm tắt
The content discusses a novel framework for legged loco-manipulation focusing on whole-body coordination through hierarchical optimization-based control. It addresses challenges in object manipulation, emphasizing the impact on robot stability and locomotion. The approach combines an online manipulation planner, pose optimization, and a whole-body loco-manipulation MPC. Experimental validation with Unitree Aliengo demonstrates successful lifting and carrying of heavy objects.
- Legged Robotics Advancements
- Integration of articulated robotic arms.
- Challenges in loco-manipulation with heavy objects.
- Hierarchical Optimization Framework
- Online manipulation planner for task-based trajectory.
- Pose optimization aligning robot trajectory with constraints.
- Whole-body loco-manipulation MPC incorporating manipulation forces.
- Control System Overview
- Three elementary sub-problems hierarchically addressed.
- Planner for Object Manipulation
- Linear MPC structure for manipulation actions.
- Pose Optimization for Coordination
- Non-linear programming problem formulation.
- Whole-Body Loco-Manipulation MPC
- Formulation addressing dynamics of heavy objects.
- Results and Experiments
- Simulation and hardware experiments validating the approach.
- Effectiveness Demonstrated:
- Importance of considering object dynamics in robot control.
- Significance of object manipulation planner and pose optimization.
Thống kê
"Our approach has been validated in simulation and hardware experiments."
"Experimental results showcase its ability to lift and carry an 8kg payload."