About

Teguh Santoso Lembono is a robotics researcher whose work spans whole-body control, motion planning, and optimization for complex robotic systems. His most impactful contribution is the first successful experimental implementation of whole-body model predictive control with state feedback on a torque-controlled humanoid robot, demonstrated on the Talos platform (47 citations). This work achieved robust balance control and target tracking under strong external perturbations, marking a significant advance in humanoid locomotion and manipulation. Lembono has also pioneered methods for multi-contact receding horizon planning, using value function approximation to guide robots in building momentum for traversing large obstacles—a critical capability for dynamic locomotion. His research extends to non-prehensile manipulation, where he combines demonstration-guided optimal control with long-term planning for underactuated systems, and to industrial robotics, including automated taping systems and riveting. Notably, he has applied variational inference and tensor train methods to solve global optimization problems in robotics, addressing the challenge of initial guess sensitivity in numerical solvers. With over 100 citations across his top papers, Lembono’s work bridges theoretical optimization, learning, and real-world robot control, making him a key figure in advancing autonomous, adaptive robotic systems.

Research Focus

Key Achievements

7
H-Index
16
Papers
142
Total Citations
9
Avg Citations/Paper
🏆 Most Cited Paper
Whole Body Model Predictive Control with a Memory of Motion: Experiments on a Torque-Controlled Talos
47 citations · 2021
📈 Most Prolific Year: 2021 (3 Papers)
🤝 Key Collaborators: 33
🏛 Institutions: Idiap Research Institute, Nanyang Technological University, École Polytechnique Fédérale de Lausanne, Foshan University, Singapore University of Technology and Design

Top Papers

  1. 1
  2. 2
  3. 3
  4. 4
  5. 5
  6. 6
  7. 7
  8. 8
  9. 9
  10. 10

Key Collaborators

Contact & Links

Available for collaboration
Content generated · 12 days ago