About Me

I am a Ph.D. student at GRASP Lab, University of Pennsylvania, advised by Prof. Nadia Figueroa. My research mainly focuses on safe and efficient motion policy learning. I’m constantly exploring how robots can generate motions to interact and collaborate with humans in various capacities. This endeavor involves methods from machine learning, control theory, and perception, aiming to bridge the gap between robot capabilities and human needs.

Before coming to Penn, I received my MS in Robotics from Northwestern University and BS in Aerospace Engineering from UIUC.

I am currently an Applied Scientist II Co-Op @ Amazon Robotics Innovation Lab

Publications

Out-of-Distribution Recovery with Object-Centric Keypoint Inverse Policy For Visuomotor Imitation Learning

George Jiayuan Gao, Tianyu Li, Nadia Figueroa

Spotlight Paper in Workshop on Lifelong Learning for Home Robots @ CoRL 2024 [website]

A vision-based object-centric recovery policy framework that guides the system back to the training distribution in OOD situations without relying on extra data

Constrained Passive Interaction Control: Leveraging Passivity and Safety for Robot Manipulators

Zhiquan Zhang, Tianyu Li, Nadia Figueroa

ICRA 2024 [website]

A novel control architecture that allows a torque-controlled robot to guarantee safety constraints such as kinematic limits, self-collisions, external collisions and singularities and is passive only when feasible

Learning Safe and Stable Motion Plans with Neural Ordinary Differential Equations

Farhad Nawaz, Tianyu Li, Nikolai Matni, Nadia Figueroa

ICRA 2024 [website]

This approach learns a motion plan with Neural Ordinary Differential Equations while guaranteeing stability and safety with Control Lyapunov Functions and Control Barrier Functions

Constraint-Aware Intent Estimation for Dynamic Human-Robot Object Co-Manipulation

Yifei Simon Shao, Tianyu Li, Shafagh Keyvanian, Pratik Chaudhari, Vijay Kumar, Nadia Figueroa

RSS 2024 [website]

A dynamical system based representation for intent estimation with constraints in human robot co-manipulation tasks

Task Generalization with Stability Guarantees via Elastic Dynamical System Motion Policies

Tianyu Li, Nadia Figueroa

CoRL 2023 [website]

A dynamical system based motion policy LfD method with stability guarantees that can generalize to new task configurations without new demonstrations

Directionality-Aware Mixture Model Parallel Sampling for Efficient Linear Parameter Varying Dynamical System Learning

Sunan Sun, Haihui Gao, Tianyu Li, Nadia Figueroa

RA-L [website]

A novel statistical model that applies the Riemannian metric on the n-sphere to efficiently blend non-Euclidean directional data with Euclidean states for learning stable, time-independent motion policies

Past Projects

Planning & Prediction with user preference via deep inverse reinforcement learning

[website]

Using maximum entropy deep IRL to learn agent preference in continuous environment path planning

Using Rethink Sawyer Robot Arm to Play Yoyo

[website]

Developed software and hardware pipeline to play yoyo with visual feedback control on the Sawyer Robot arm

Real-time KL-Ergodic Distribution-based Model Predictive Control

[website]

Model predictive control with the ability to match or avoid distributions

EKF-SLAM with Machine Learning

[website]

Implementation of landmark-based EKF-SLAM with unsupervised learning and unknown data association using ROS in C++ from scratch.

Data-driven Receding Horizon Control with the Koopman Operator

[website]

Performed receding horizon control using data-driven approach with small data in a continuous space