Garrett Thomas

Papers

Total Citations

H-Index

About

Garrett Thomas is a robotics researcher whose work focuses on integrating machine learning with robotic manipulation, particularly for industrial assembly tasks. His key research areas include autonomous assembly, model predictive control (MPC), and learning from demonstration. Thomas's major contribution lies in developing methods that enable robots to acquire complex contact-rich manipulation skills without relying on classical control approaches. In his 2018 paper "Learning Robotic Assembly from CAD," he pioneered a framework for robots to learn assembly tasks directly from computer-aided design (CAD) models, addressing a critical gap in modern manufacturing. This work has garnered 10 citations and represents a significant step toward automating industrial processes. His earlier 2017 paper "Learning from the hindsight plan — Episodic MPC improvement" (7 citations) introduced an innovative approach to improving model predictive control by learning from past planning episodes, enhancing both computational efficiency and robustness. Thomas's research bridges the gap between theoretical control methods and practical robotic applications, making him a notable figure in the field of learning-based robotics. His work continues to influence how robots can autonomously acquire and refine manipulation skills for real-world manufacturing environments.