Description

This thesis aims to bridge the disparity between high-level instructions in large AI models for robotics, and low-level robot execution. This will be tackled by building a common features framework for variable impedance controllers to understand different kinds of inputs, such as language, proprioceptive information
(vision, forces, etc.) and user preferences. This will allow for a context aware definition of such controllers and their reference, taking user preferences into account naturally.

The work is under the scope of the following projects:

• DROVIC: Data-driven Robot Variable Impedance Control with Context Awareness (web)