Seyed Mohammad Khansari-Zadeh
École Polytechnique Fédérale de Lausanne, Stanford University
Papers
10
Total Citations
1,620
H-Index
8
About
Seyed Mohammad Khansari-Zadeh is a leading researcher in robot learning and control, best known for pioneering methods that enable robots to learn stable, human-like motions from demonstrations. His core contributions lie at the intersection of dynamical systems theory and imitation learning, where he has developed frameworks to ensure that learned robot movements are both globally asymptotically stable and robust to perturbations. His most influential work, "Learning Stable Nonlinear Dynamical Systems With Gaussian Mixture Models" (2011), has amassed over 764 citations and introduced a foundational approach for encoding discrete robot motions as time-invariant dynamical systems. He further advanced the field with key contributions to real-time obstacle avoidance (259 citations) and learning control Lyapunov functions for stable reaching motions (224 citations). His iterative algorithm BM (2010) and nonlinear programming methods for imitation learning have set standards for stability guarantees in robot motion generation. Notably, his work extends beyond theory to practical applications, such as teaching robots to play minigolf, demonstrating how dynamical system-based approaches can handle complex, interactive tasks. Khansari-Zadeh’s research has profoundly influenced how robots learn and adapt motions safely and efficiently.
Research Focus
Key Achievements
Top Papers
- 1Learning Stable Nonlinear Dynamical Systems With Gaussian Mixture Models764 citations · 2011
- 2A dynamical system approach to realtime obstacle avoidance259 citations · 2012
- 3
- 4Learning Non-linear Multivariate Dynamics of Motion in Robotic Manipulators116 citations · 2010
- 5
- 6
- 7
- 8Learning to Play Minigolf: A Dynamical System-Based Approach26 citations · 2012
- 9
- 10