Papers
7
Total Citations
38
H-Index
2
About
Yann Chevaleyre is a leading researcher in robotics and artificial intelligence, with a primary focus on bridging the gap between simulation and real-world deployment for autonomous systems. His most impactful work centers on humanoid robotics and exoskeleton control, where he has pioneered the use of reinforcement learning for reactive stepping and push recovery—critical capabilities for bipedal robots operating in unpredictable environments. His 2022 paper on the Atalante exoskeleton (15 citations) demonstrates state-of-the-art methods for standing balance recovery, directly addressing the notorious "reality gap" that plagues sim-to-real transfer. Beyond locomotion, Chevaleyre has made foundational contributions to robot perception and learning, including meta-learning approaches for grounding symbols from visual data (2003, 13 citations) and wrapper-based methods for adaptive object detection. His work also extends to multi-agent systems, tackling challenges in swarm re-localization and patrol problems. By combining trajectory optimization with function approximation, he has developed efficient online planning algorithms that enable real-time decision-making for autonomous robots. Chevaleyre’s research consistently emphasizes practical, hardware-validated solutions, making him a key figure in advancing the real-world capabilities of legged robots and multi-robot teams.
Research Focus
Key Achievements
Top Papers
- 1
- 2A meta-learning approach to ground symbols from visual percepts13 citations · 2003
- 3
- 4Wrapper for object detection in an autonomous mobile robot2 citations · 2003
- 5The robot swarm re-localization problem2 citations · 2009
- 6A Wrapper-Based Approach to Robot Learning Concepts from Images2 citations · 2002
- 7Le problème multi-agents de la patrouille2 citations · 2011