RESEARCH

Robot Systems

We aim at providing practical and effective control solutions to the real systems such as robot manipulators, humanoids, and so on. Through such solutions, we are expected to achieve stability and optimality of the real control systems. The detailed research topics tackled in the Control Theory & Robotics, Learning Laboratory are as follows (but not limited to):

• Robot manipulators

         1)   Optimal control for dexterous manipulation 

         2)   Robust control for stable human-robot interaction 

• Quadruped Robots

         1)   Safety recovery of quadruped robots for various disturbance

         2)   Walking control of quadruped robots for various disturbance 

• Mobile Robots

         1)   Safe navigation of Mobile robots for dynamic disturbance 

• Cyber-physical systems

         1)   Stable tele-operation systems (e.g., surgical robots)

         2)   Cyber security for adversarial attacks 

Learning-Based Control

We are interested in developing learning-based control algorithms for uncertain, nonlinear, and complex dynamical systems. By integrating control theory, optimization, and machine learning, we aim to design intelligent controllers that can learn from data while maintaining stability, safety, and robustness. The detailed research topics considered in the Control Theory & Robotics, and Learning Laboratory are as follows (but not limited to): 

• Learning-based control for dynamical systems 

         1)   Safe and robust learning-based control under uncertainties and disturbances 

         2)   Data-driven controller design with stability guarantees 

• Reinforcement learning and optimal control 

         1)   Reinforcement learning-based control for robotic systems 

         2)  Learning-based optimal control under constraints