Neurosymbolic planning with an UR5e-Robot

  • Subject:Robotics, Neurosymbolic Planning, Large Language Models
  • Type:Bachelor or Master Thesis
  • Supervisor:

    Lukas Kinder

  • Add on:

    Our lab is equipped with a UR5e robot that can be controlled via external code. This makes it possible to generate planning sequences for the robot using a Large Language Model (LLM). However, LLMs alone can struggle to control a robot when tasks require many precise movements or substantial computation. In this thesis, we will explore a neuro-symbolic approach to improve LLM-based robot control. The project can be done as either a Bachelor or Master Thesis. If interested, please email me at lukas.kinder@kit.edu.