Brent Griffin
Papers
10
Total Citations
375
H-Index
6
About
Brent Griffin is a robotics researcher whose work spans two compelling domains: bipedal locomotion control and robot perception for manipulation. He is perhaps best known for his foundational contributions to underactuated 3D bipedal walking, where he helped bridge the critical gap between planar theoretical models and real-world three-dimensional implementation. His landmark 2016 paper on 2D-to-3D gait design (110 citations) and his experimental demonstrations with the MARLO robot (78 citations) established him as a key contributor to dynamic legged locomotion. Central to this work is his development and refinement of virtual constraints and nonholonomic control frameworks, which enable robots to walk robustly over uneven terrain — a persistent challenge in the field. His gait optimization methods, which systematically account for terrain uncertainty, have been widely adopted by the community. In later work, Griffin expanded into robot perception and manipulation, exploring object detection, video segmentation, and self-supervised learning to enable robots to recognize and interact with novel objects in unstructured environments. With over 370 cumulative citations, his research reflects a rare breadth — from elegant mathematical control theory to practical, data-driven robot autonomy.
Research Focus
Key Achievements
Top Papers
- 1
- 2Preliminary walking experiments with underactuated 3D bipedal robot MARLO78 citations · 2014
- 3
- 4
- 5Nonholonomic virtual constraints for dynamic walking36 citations · 2015
- 6
- 7Mobile Robot Manipulation using Pure Object Detection6 citations · 2023
- 8
- 9Self-Supervised Robot In-hand Object Learning.3 citations · 2019
- 10