Tianzhi Li 李天志
 |
Incoming Postdoctoral Fellow (Shun Hing Fellow) |
About me
I am an incoming Shun Hing Postdoctoral Fellow at the Chinese University of Hong Kong (CUHK), working with Prof. Zhongyu Li. I received my Ph.D. in Mechanics & Control from Peking University (PKU) in June 2026, working with Prof. Jinzhi Wang and Prof. Zhisheng Duan. From Dec-2024 to Dec-2025, I was a visiting researcher at Nanyang Technological University (NTU), Singapore, advised by Prof. François Gay-Balmaz. Prior to that, I received my B.Sc. in Mathematics from Beijing Institute of Technology (BIT) in 2021 under the supervision of Prof. Donghua Shi.
Research Interest
I develop geometric methods with structure-preserving properties for robot learning & control.
Geometric Mechanics (especially stochastic & nonholonomic systems)
Geometric Safety-Critical Control (for humanoid robots)
Geometric Robot Learning for Humanoids
Structure-Preserving Discretization (variational integrators)
Selected Research
 |
Variational Principle for Stochastic Nonholonomic Systems Part II: Stochastic Nonholonomic Integrator International Conference on Geometric Science of Information (GSI 2025). Lecture Notes in Computer Science, vol 16034, pp. 225-233. Springer, Cham, 2026. We propose stochastic discrete variational principles for stochastic nonholonomic systems and construct the associated stochastic nonholonomic integrator. |
 |
Variational Unscented Kalman Filter on Matrix Lie Groups Automatica, 172: 111995, 2025 (Regular Paper). We propose a family of computationally efficient unscented Kalman filters (UKF-Vs) for mechanical systems on matrix Lie groups. |
 |
Reduced Dynamics and Geometric Optimal Control of Nonequilibrium Thermodynamics: Gaussian Case Automatica, 164: 111626, 2024 (Regular Paper). We study the geometric structures of n-DOF Gaussian distributions, and we propose a geometric optimal control algorithm for minimum-energy optimal control problem of Gaussian distributions. |
 |
Physics-Informed Gaussian Process Learning on Lie Groups Journal of Guidance, Control, and Dynamics, 48 (11), pp. 2654-2662, 2025. We introduce a physics-informed Gaussian process learning method for mechanical systems on Lie groups and on a special class of homogeneous manifolds based on discrete mechanics theory. |
News
Jun 2026: Today (June 30) I attended a wonderful graduation ceremony with my family at AMR, PKU. So excited to enjoy this happy moment!
May 2026: I am so happy that I defended my PhD thesis on 13 May and received the honor of Outstanding Graduate of Beijing.
Aug 2025: Our paper Structure-Preserving Unscented Kalman Filter for Planar Mobile Robots has been accepted to IEEE Control Systems Letters (IEEE L-CSS).
Aug 2025: Our paper Physics-Informed Gaussian Process Learning on Lie Groups has been accepted to Journal of Guidance, Control, and Dynamics (JGCD).
Jul 2025: Our two papers Variational Principle for Stochastic Nonholonomic Systems Part I & Part II (joint work with Professors François Gay-Balmaz, Donghua Shi, and Jinzhi Wang) have been accepted by the International Conference on Geometric Science of Information (GSI 2025).
Sep 2024: Our paper Variational Unscented Kalman Filter on Matrix Lie Groups has been accepted to Automatica.
Feb 2024: Our paper Reduced Dynamics and Geometric Optimal Control of Nonequilibrium Thermodynamics: Gaussian Case has been accepted by Automatica.