Robin Deits
Massachusetts Institute of Technology, Vassar College, Battelle
Papers
22
Total Citations
2,114
H-Index
16
About
Robin Deits is a leading researcher in legged robotics, whose work bridges perception, planning, and control to enable humanoid robots to navigate complex, real-world environments. His most impactful contributions center on optimization-based locomotion, where he developed novel methods for footstep planning on uneven terrain using mixed-integer convex optimization—a technique that elegantly handles obstacle avoidance and kinematic constraints. This foundational work, cited over 280 times, directly enabled the Atlas humanoid robot to achieve robust, dynamic walking. Deits also pioneered the computation of large convex regions of obstacle-free space via semidefinite programming, a widely-adopted approach (over 220 citations) for safe motion planning. His broader portfolio includes closed-form solutions for real-time gait stabilization, stereo-vision-based locomotion over rough terrain, and information-theoretic dialog systems for human-robot communication. Notably, his early-career work on the bio-inspired RoboClam—analyzing burrowing drag reduction in razor clams—demonstrates a rare versatility, bridging biology and robotics. With over 1,800 total citations, Deits’ research has shaped modern humanoid control, from the DARPA Robotics Challenge to ongoing advances in whole-body autonomy.
Research Focus
Key Achievements
Top Papers
- 1
- 2Footstep planning on uneven terrain with mixed-integer convex optimization283 citations · 2014
- 3
- 4An Architecture for Online Affordance‐based Perception and Whole‐body Planning140 citations · 2014
- 5
- 6
- 7Continuous humanoid locomotion over uneven terrain using stereo fusion75 citations · 2015
- 8
- 9Clarifying Commands with Information-Theoretic Human-Robot Dialog59 citations · 2013
- 10Toward Information Theoretic Human-Robot Dialog50 citations · 2012