Proposing GOTT, a reusable object grasping technology for multiple robot hands
- What
- To enable multi-fingered robot hands, similar to human hands, to stably grasp and move objects in desired directions, the authors proposed GOTT, a contact technology that can be commonly used across different robots and hand shapes.
- Key result
- In simulations and real-world robot experiments, GOTT established stable contacts across various objects, arm-hand platforms, and unseen hand shapes, and consistently improved overall task success rates compared to pre-planned grasping methods.
- Why it matters
- This study demonstrates the potential to reuse the same technology even when robot hand types change by using a common motion component not dependent on a specific robot. However, it is difficult to judge the commercialization level or real industrial impact from the abstract alone.
Background
Large-scale AI models and human demonstration data provide rich information on how to handle objects, but translating this into actual movements for multi-fingered robots remains challenging. In particular, there is a lack of reusable primitive motions that allow robot hands to make stable contacts across various tasks and robots.
Method
The researchers proposed the GOTT framework, which consists of bringing the hand near the object, stably grasping it using a common contact technology, and moving it along the object's target pose. The object's motion trajectory and approach instructions can be received from various methods such as look-ahead planning, external models, or human demonstrations, while the core contact technology and tracking controller are kept unchanged.
Main results
Simulations and real-world robot experiments report robust contact capabilities for various objects and robot platforms. It was applied to hand shapes not used during training, and consistently improved overall task success rates compared to open-loop simple grasping execution.
Limitations & open points
The abstract lacks specific success rate figures, the number of tested objects and robots, and failure rate percentages, making it difficult to quantitatively verify the degree of performance. Furthermore, the abstract does not confirm whether it operates at the same level in complex real-world environments.
Glossary
- multi-fingered hand — A robot hand equipped with multiple fingers like a human hand, capable of delicately grasping and rotating objects.
- cross-embodiment — Means that the technology can be applied in the same way even if the shapes of the robot arms or hands differ.
- foundation model — A large-scale AI model pre-trained on massive data so it can be widely used across various tasks.
- closed-loop control — A control method where the robot continuously checks its current state with sensors and modifies its movements.
- open-loop execution — A method of executing a predetermined sequence of motions without sensor feedback.
- end-to-end task success — The rate at which the entire process from approach to object manipulation is completed from start to finish.
This summary was written by AI from the paper's abstract and reviewed by the editor. Summaries on this page were translated by AI from the editor-reviewed Korean version. Check the original for exact figures and context.
This page compiles published papers, official announcements and press reports for information only. Paper summaries are written by AI from the abstract and reviewed by the editor; they may differ from the original, so check the paper for anything important. Nothing here is a recommendation to buy or sell any security or investment advice. Career moves cite official announcements or major press reports; some career records come from Wikidata (CC0). Summaries on this page were translated by AI from the editor-reviewed Korean version.
