Agentic Language-Grounded Adaptive Robotic Assembly

Nicholas Cote, Jaimyn Drake, Sachin ChittaRobo-3Dvlm @ CVPR 2025 · OpenReview

Abstract

High-mix, low-volume manufacturing requires systems that can adapt to changing parts and processes. We present an agentic, language-grounded approach for robotic assembly that employs a team of Large Language Models and Vision-Language-Action Models to adapt to parts of similar kind and perform corrective actions during assembly. Given a high-precision wheel-on-axle insertion task, we demonstrate that our agentic approach outperforms a single-model and generalizes to out-of-distribution parts and grasps without changing the nominal assembly process.
BibTeX
@inproceedings{cote2025agentic,
  author = {Nicholas Cote and Jaimyn Drake and Sachin Chitta},
  title = {{Agentic Language-Grounded Adaptive Robotic Assembly}},
  booktitle = {Robo-3Dvlm @ CVPR},
  year = {2025},
  url = {https://openreview.net/forum?id=T5D0SDpw16},
}

Loading related content...

Figure