CushionCatch: Compliant Catching Mechanism for Mobile Manipulators via Combined Optimization and Learning

📅 2024-09-23
🏛️ arXiv.org
📈 Citations: 0
Influential: 0
📄 PDF

career value

221K/year
🤖 AI Summary
This work addresses the challenge of achieving compliant buffering during high-speed aerial object capture by mobile manipulators. Methodologically, it introduces an end-to-end compliant grasping framework integrating optimization and learning: (1) joint optimization of capture-point planning and pre-grasp motion; (2) a position-encoded LSTM (P-LSTM) to model spatiotemporal compliance policies, enabling effective imitation of human demonstrations; and (3) real-time collision-aware post-grasp trajectory tracking. To the best of our knowledge, this is the first framework achieving closed-loop compliant capture—from planning to execution—on mobile manipulators. Evaluations show a 98.70% success rate in simulation and 92.59% on physical hardware, with a 28.7% reduction in impact torque. The implementation is publicly available.

Technology Category

Application Category

📝 Abstract
Catching flying objects with a cushioning process is a skill commonly performed by humans, yet it remains a significant challenge for robots. In this paper, we present a framework that combines optimization and learning to achieve compliant catching on mobile manipulators (CCMM). First, we propose a high-level capture planner for mobile manipulators (MM) that calculates the optimal capture point and joint configuration. Next, the pre-catching (PRC) planner ensures the robot reaches the target joint configuration as quickly as possible. To learn compliant catching strategies, we propose a network that leverages the strengths of LSTM for capturing temporal dependencies and positional encoding for spatial context (P-LSTM). This network is designed to effectively learn compliant strategies from human demonstrations. Following this, the post-catching (POC) planner tracks the compliant sequence output by the P-LSTM while avoiding potential collisions due to structural differences between humans and robots. We validate the CCMM framework through both simulated and real-world ball-catching scenarios, achieving a success rate of 98.70% in simulation, 92.59% in real-world tests, and a 28.7% reduction in impact torques. The open source code will be released for the reference of the community.
Problem

Research questions and friction points this paper is trying to address.

Develop compliant catching for mobile manipulators using optimization and learning.
Design a network combining LSTM and positional encoding for compliant strategies.
Validate framework in simulations and real-world with high success rates.
Innovation

Methods, ideas, or system contributions that make the work stand out.

Combines optimization and learning for compliant catching
Uses P-LSTM network for learning from human demonstrations
Achieves high success rates in simulation and real-world
B
Bingjie Chen
Center for Artificial Intelligence and Robotics, Shenzhen International Graduate School, Tsinghua University, Shenzhen 518055, China
K
Keyu Fan
Center for Artificial Intelligence and Robotics, Shenzhen International Graduate School, Tsinghua University, Shenzhen 518055, China
H
Houde Liu
Center for Artificial Intelligence and Robotics, Shenzhen International Graduate School, Tsinghua University, Shenzhen 518055, China
C
Chongkun Xia
School of Advanced Manufacturing, Sun Yat-Sen University, Shenzhen 518055, China
L
Liang Han
School of Electrical and Automation Engineering, Hefei University of Technology, Hefei 230009, China
B
Bin Liang
Navigation and Control Research Center, Department of Automation, Tsinghua University, Beijing 100084, China