Title: Trajectory Modeling using Generative Approaches for Scheduling, Planning, and Multi-Agent Systems
Date: Monday, August 19th
Time: 1:00pm-3:00pm Eastern time
Location: Klaus 1116E
Zoom link: https://gatech.zoom.us/my/seanye
Sean Ye
Robotics PhD Candidate
School of Aerospace Engineering
Georgia Institute of Technology
Committee:
Dr. Matthew Gombolay (Advisor) - School of Interactive Computing, Georgia Institute of Technology
Dr. Karen Feigh- School of Aerospace Engineering, Georgia Institute of Technology
Dr. Harish Ravichandar- School of Interactive Computing, Georgia Institute of Technology
Dr. David Sidoti- Distinguished Member of R&D, US Naval Research Laboratory
Dr. Andreea Bobu- Department of Aeronautics and Astronautics, MIT
Abstract:
Trajectory prediction and generation are crucial for autonomous robots navigating dynamic environments. While previous research has often focused on either prediction or generation, my dissertation explores how deep generative modeling can address both aspects within a unified framework. I discuss applications in various robotics domains, including long-horizon planning, adversarial tracking, trajectory forecasting, and agile robot control. I present several methods to enhance generative models for modeling multi-agent teams, integrating different constraints, and reducing sample-time complexity to enable application on real robots. Finally, I present work evaluating large language model's capabilities for increasing usability and reducing workload for scheduling and motion planning. Overall, this thesis aims to demonstrate the versatility and effectiveness of deep generative modeling in advancing robot autonomy.