Robotics Seminar

Content

Abstract

Robotics is rapidly evolving to address challenges in dynamic, human-centric environments. However, enabling robots to collaborate effectively with humans and adapt to diverse individuals, tasks, conditions, and preferences remains a significant challenge. This talk will focus on two key research areas addressing these issues: human multi-robot systems and robot learning. The first part of the talk will explore human multi-robot systems, highlighting strategies for effective collaboration between humans and robot teams. I will introduce a dataset developed to advance research in this field and discuss our work on task allocation, including initial assignment, which accounts for diverse team capabilities and attributes, and dynamic allocation, which enables real-time adaptation based on human state and performance. The second part will focus on robot learning, aiming to enhance personalization and usability for diverse users. I will present our research on preference-based learning for personalized interactions, learning-from-demonstration frameworks where robots learn directly from human demonstrations, and large language model-driven reasoning and learning techniques that enable multi-robot formation using only high-level contextual instructions from users.