Technical Capabilities

Technical Capabilities

Through CLBF we achieve safety and stability, through transfer learning we ensure generalization and transferability, and through physical-information priors we improve sample efficiency — these three form the theoretical cornerstone for the deployable "cerebellum" of embodied intelligence.

Academic Achievements

01

Optimal Control of Nonlinear Systems Based on Stable and Safe Reinforcement Learning

AIChE Journal (top chemical engineering journal)

02

Transfer Reinforcement Learning Control Algorithm Design and Its Generalization Performance Analysis

IEEE Transactions on Cybernetics

03

Physics-Informed Reinforcement Learning Optimal Control Algorithm

Theoretical framework for embedding physical priors into policy learning

04

Cerebellum Key Algorithm Optimization — Fine Manipulation and Load Mutation

TDSMC (time-delay SMC) sliding mode control