Authors: Jiaqiang Ye Zhu, Carla Gomez Cano, David Vazquez Bermudez, Michal Drozdzal

Published on: February 07, 2024

Impact Score: 8.22

Arxiv code: Arxiv:2402.05188

Summary

• What is new: The introduction of InCoRo, a system that extends previous work to enable robots to operate in dynamic environments using LLMs for reasoning, surpassing prior art in terms of success rate.
• Why this is important: The difficulty in empowering robots with robust reasoning capabilities to execute complex tasks in dynamic settings.
• What the research proposes: InCoRo, a system combining an LLM controller, scene understanding, and a robotic unit to analyze environmental changes and adapt execution commands in real-time.
• Results: In static environments, InCoRo showed improved success rates over previous methods. In dynamic environments, it established new benchmarks for performance using SCARA and DELTA robotic units.

Technical Details

Technological frameworks used: Classical robotic feedback loop

Models used: Off-the-shelf LLM model

Data used: nan

Potential Impact

Robotics, automation industries, and companies involved in manufacturing, logistics, and potentially service sectors could be disrupted or benefit.

Want to implement this idea in a business?

We have generated a startup concept here: AdaptiBot Solutions.