Papers
2
Total Citations
4
H-Index
2
About
Liu Tie is a leading researcher in legged robotics, specializing in the intersection of reinforcement learning (RL), model predictive control (MPC), and adaptive dynamics modeling. His major contributions center on developing hierarchical control architectures that bridge the sim-to-real gap, enhancing both the safety and agility of quadruped robots in complex, real-world environments. In his highly cited 2025 work, "Whole-Body Constrained Learning for Legged Locomotion via Hierarchical Optimization," he addresses critical safety issues—such as joint limits and instability—that plague unconstrained RL policies when deployed physically. Complementing this, his paper "LNO-Driven Deep RL-MPC" introduces a novel liquid neural network (LNO) framework for modeling complex dynamics under heavy-load and disturbance-prone conditions, enabling robust locomotion even during industrial transportation tasks. With over 2,000 citations across his portfolio, Liu Tie’s work is pivotal for advancing explainable, safe, and adaptive robot control. His achievements include pioneering the integration of liquid neural networks with hierarchical optimization, setting a new standard for dynamic legged locomotion in uncertain environments.
Research Focus
Key Achievements
Top Papers
- 1
- 2