Russ Tedrake
McGovern Institute for Brain Research, Massachusetts Institute of Technology, Moscow Institute of Thermal Technology, MIT-Harvard Center for Ultracold Atoms, Artificial Intelligence in Medicine (Canada), Harvard University, IIT@MIT, Vassar College, Toyota Research Institute, Intel (United States)
Papers
118
Total Citations
10,563
H-Index
51
About
Russ Tedrake is a pioneering roboticist whose research spans legged locomotion, trajectory optimization, motion planning, and increasingly, robot learning. Best known for his foundational work on bipedal walking, his 2005 paper on passive-dynamic walkers — now approaching 1,900 citations — revealed how humanlike locomotion can emerge from purely mechanical systems, reshaping how researchers model and design walking robots. This insight seeded decades of subsequent work, including reinforcement learning applied directly to physical 3D bipeds and robust dynamic walking over uneven terrain. Tedrake's contributions to humanoid robotics reached a landmark with his team's optimization-based control framework for Boston Dynamics' Atlas robot (814 citations), demonstrating that whole-body planning with centroidal dynamics and full kinematics could make complex humanoids practically deployable. His pioneering trajectory optimization methods for rigid-body contact (608 citations) and mixed-integer footstep planning remain cornerstones of the field. More recently, Tedrake has pushed into robot learning and manipulation, with his work on Diffusion Policy (338 citations) representing a striking pivot — applying generative diffusion models to visuomotor control and earning rapid recognition from the broader machine learning community. Across fields ranging from controls theory to modern deep learning, Tedrake's research consistently bridges elegant theory with real-world robotic performance.
Research Focus
Key Achievements
Top Papers
- 1Efficient Bipedal Robots Based on Passive-Dynamic Walkers1,878 citations · 2005
- 2
- 3A direct method for trajectory optimization of rigid bodies through contact608 citations · 2013
- 4Whole-body motion planning with centroidal dynamics and full kinematics418 citations · 2014
- 5Diffusion policy: Visuomotor policy learning via action diffusion338 citations · 2024
- 6Belief space planning assuming maximum likelihood observations308 citations · 2010
- 7Footstep planning on uneven terrain with mixed-integer convex optimization283 citations · 2014
- 8Stochastic policy gradient reinforcement learning on a simple 3D biped270 citations · 2005
- 9
- 10Stable dynamic walking over uneven terrain212 citations · 2011