Caelan Reed Garrett
Massachusetts Institute of Technology, Nvidia (United Kingdom), Nvidia (United States)
Papers
26
Total Citations
696
H-Index
11
About
Caelan Reed Garrett is a leading researcher in robot autonomy, whose work fundamentally bridges the gap between high-level symbolic reasoning and low-level physical execution. His core research areas are task and motion planning (TAMP), robot manipulation, and the integration of learning with classical planning. Garrett’s most significant contribution is the development of the PDDLStream framework (174 citations), which provides a principled method for integrating symbolic planners with continuous blackbox samplers, enabling robots to reason about complex geometric and kinematic constraints. He has also pioneered massively parallel motion generation with CuRobo (82 citations), demonstrating that global optimization on GPUs can solve collision-free motion problems at unprecedented speeds. His work on learning compositional models of robot skills (82 citations) and sampling-based methods for factored TAMP (81 citations) has been instrumental in enabling robots to solve long-horizon manipulation problems with unknown objects. Garrett’s research has been recognized for its practical impact on multi-arm assembly systems and its forward-looking integration with vision-language models, making him a key figure in the next generation of autonomous robotic systems.
Research Focus
Key Achievements
Top Papers
- 1
- 2Learning compositional models of robot skills for task and motion planning82 citations · 2021
- 3CuRobo: Parallelized Collision-Free Robot Motion Generation82 citations · 2023
- 4Sampling-based methods for factored task and motion planning81 citations · 2018
- 5
- 6
- 7
- 8Cooperative Task and Motion Planning for Multi-Arm Assembly Systems20 citations · 2022
- 9Integrated Task and Motion Planning17 citations · 2021
- 10Guiding Long-Horizon Task and Motion Planning with Vision Language Models16 citations · 2025