Understanding and Translation of Everyday Verbal Instructions for a Robotic Agent
Large Language Models (LLMs) are currently a very popular research topic. Our goal is to transfer current developments to real robotic systems as this has a big potential in terms of a more intuitive way to interact with an embodied AI agent.
We will investigate this complex topic in smaller sub-problems. One of these subprojects is the following topic: How can we convert arbitrary text instructions (possibly from a speech-to-text module) into a valid API instruction for an existing or to be extended robot control software. The investigations are to be supported by an example application in ROS2 (Robot Operating System).
Advisor:
- Sven Lange, sven.lange@…
Requirements:
- Basic knowledge in Linux/Ubuntu would be good.