Impact-Aware Dual-Arm Manipulation
James Hermus, Michael Bosongo Bombile, Jari J. van Steen, Elise Jeandupeux, Ahmed Zermane, Alessandro Melone, Mario Troebinger, Abdeldjallil Naceri, Claude Lacoursière, Stijn de Looijer, Sami Haddadin, Abderrahmane Kheddar, Alessandro Saccon, Aude Billard
AI summary
Problem
Traditional robotic depalletization relies on slow, energy-intensive quasi-static interactions that cannot match human speed or adaptability, creating a bottleneck in fast-growing e-commerce logistics.
Approach
The framework integrates dynamical systems for intentional impact motion generation, reference spreading to smooth post-impact velocity jumps, and quadratic programming to enforce hardware constraints during dynamic grabbing and tossing.
Key results
- 29% reduction in average task execution time
- 35% decrease in energy consumption
- Robust dual-arm coordination via dynamical systems
- Contact state estimation without external force sensors
Why it matters
This approach provides a scalable, high-throughput automation solution for warehouse logistics that bridges the performance gap between human workers and traditional robotic systems.
Abstract
No abstract on file.