The future of home robotics is rapidly approaching. When it comes to non-professional use cases, it is not enough for a robot to simply be able to achieve a fully specified goal. In typical settings, the ,nds given to a robot will usually be abstract and ambiguous.
The goal of this project is to study and build a framework for robot understanding of goal-related information. This framework will query from robots’ observation model, transform multi-modal information into structured representation, do goal inference for human and action recommendation for robots based on the structured information. We will demonstrate the system’s ability of physical-world inference and recommendation in a domestic setup.

“I’m participating in SuperUROP to get support and supervision from my lab to build the AI-based robotics system that I’m very excited about.
Through the SuperUROP, I hope I can make some technical contribution to the research domain, and also get myself prepared with domain knowledge and technical skills to continue working towards this direction after school.”