Event

Proposal of an asynchronous upper-body trajectory tracking framework for humanoid robots

1 reports · 1 sources

Yumeng Liu et al. propose an asynchronous upper-body task-space trajectory tracking framework for humanoid robots, addressing the temporal asynchrony between sparse, low-frequency trajectories output by high-level planners and the high-frequency operation of whole-body controllers. The student policy is initialized via teacher-student distillation, conditioned on the fully cached future trajectory and execution-time index, and trained with a sliding-window global reward to reduce frame drift; in the post-task training stage, an MPC module completes sparse references into floating-base and upper-body guidance. The paper reports that in simulation and Unitree G1 hardware experiments, the method achieves better tracking under low update rates, outperforms synchronous and decoupled baselines, and adapts more safely to out-of-distribution motions.