Georgia Institute of Technology
Trajectory Modeling using Generative Approaches for Scheduling, Planning, and Multi-Agent Systems
Abstract
dc:description.abstractThis dissertation explores the application of deep generative models to trajectory forecasting and planning in robotics. Traditional approaches often rely on known motion models or complex optimization techniques, which can be limiting in dynamic, partially observable environments. The research presents novel methods that leverage advancements in generative machine learning to address these challenges. The contributions include: (1) development of single-target and multi-target adversarial tracking models, (2) creation of a flow-matching framework for trajectory prediction and generation, (3) a multi-agent reinforcement learning algorithm for large, sparse partially observable domains, (4) a kinematically constrained gradient guidance method for learning diverse, agile tasks on physical robots, and (5) integration of pre-trained large language models to enhance usability in scheduling and motion planning tasks. These approaches demonstrate improved accuracy in trajectory prediction, increased performance in multi-agent scenarios, and enhanced user experience in human-robot interaction. This work bridges classical robotics techniques with modern machine learning, advancing the field towards more adaptive and capable autonomous systems.
Degree
thesis:*- Level thesis:degree_level
- Doctoral
- Department dc:contributor.department
- Aerospace Engineering
- Grantor dc:publisher
- Georgia Institute of Technology
- Year dc:date.issued
- 2024
Author and committee
dc:creator, dc:contributor.*- Author dc:creator
-
- Ye, Sean Charles
- Advisor dc:contributor.advisor
-
- Gombolay, Matthew
- Committee members dc:contributor.committeemember
-
- Feigh, Karen
- Ravichandar, Harish
- Bobu, Andreea
- Sidoti, David
- Gombolay, Matthew
Subjects
dc:subject × 1Rights
- Language dc:language.iso
- en_US
Identifiers
dc:identifier.*- Handle dc:identifier.uri
- https://hdl.handle.net/1853/77761