Authors: Yannick Burkhardt, Qian Feng, Karan Sharma, Zhaopeng Chen, Alois Knoll

Published on: October 27, 2023

Impact Score: 8.22

Arxiv code: Arxiv:2310.17923

Summary

• What is new: First method to achieve dynamic multi-fingered grasping for unknown objects.
• Why this is important: Existing autonomous robots struggle to reliably grasp a wide variety of objects in unstructured environments with simple end-effectors.
• What the research proposes: Integrating a five-finger hand with visual servo control and a deep learning-based generative grasping network to enhance grasping variety and adapt to dynamic disturbances.
• Results: Confirmed capability to reliably grasp unknown dynamic target objects with real hardware experiments.

Technical Details

Technological frameworks used: Deep learning

Models used: Generative grasping network

Data used: Visual sensor data

Potential Impact

Robotics manufacturing, automation industries, and companies integrating autonomous systems in dynamic environments.

Want to implement this idea in a business?

We have generated a startup concept here: FlexiGrip.