LooperMuscle is introduced, a composed expert policy learning framework that restores tracking quality while preserving high training efficiency, and substantially outperforms vanilla FastSAC in motion tracking accuracy while requiring far less wall-clock time than PPO.
Abstract
FastSAC-style methods significantly reduce humanoid motion training time but often suffer from notable performance degradation compared with PPO in whole-body tracking tasks. We target this speed-performance gap by introducing LooperMuscle, a composed expert policy learning framework that restores tracking quality while preserving high training efficiency. LooperMuscle combines a semantically structured mixture-of-experts actor, an expert-aware distributional critic, and contribution-routed replay with deferred curriculum scheduling. These three components form a closed training loop in which expert contributions guide data routing, routed data shape value learning, and value gradients in turn refine expert specialization. Empirically, our approach substantially outperforms vanilla FastSAC in motion tracking accuracy while requiring far less wall-clock time than PPO: where FastSAC trains in about 15 minutes but underperforms, and PPO achieves stronger results but requires about 6 hours, LooperMuscle recovers a substantial fraction of the remaining gap to PPO in roughly 45 minutes of simulation training, delivering practical efficiency for rapid policy iteration. The code will be released to benefit the research community at https://loopermuscle.github.io/.
StableMimic is presented, a unified tracker trained beyond the nominal tracking distribution that achieves the lowest errors on all four tracking metrics among five methods and attains the lowest values on six of seven post-fall motion and load measures, supporting improved interaction safety under this protocol.
Weihao Wu, Mingzhe Huang, Ruofei Liu et al.· 0 citations
Athena-WBC is proposed, a compact teacher-student pipeline with capability-aligned policy experts for long-tail humanoid whole-body control that shows improved recovery of training-set long-tail motions and better held-out tracking than a strong SONIC-recipe baseline, using only a small number of experts.
Yuanchang Jiang, Ningyuan Zhang, Xicun Yang et al.· 1 citation
Humanoid general motion tracking requires policies that can follow diverse whole-body references while maintaining balance. Building such policies involves many practical design choices, and their individual effects are often hard to assess. We address this issue with an empirical study of common modeling and training factors used in recent humanoid motion-imitation pipelines. To make the study controlled and reproducible, we developed YAHMP, an open-source modular framework for training, evaluating, and deploying whole-body motion tracking policies on the Unitree G1. Within YAHMP, we define a nominal configuration and compare variants that differ in motion-command representation, observation history, action representation, actuation profile, hand-force randomization during training, and training approach. We evaluate the resulting policies on a test set of retargeted human motions and compare the nominal policy with TWIST2 as an external baseline trained on the same motion set. The results distinguish choices with clear tracking effects from choices that mainly change actuation effort, training complexity, or physical interaction capability. Finally, we deploy YAHMP policies zero-shot on the real Unitree G1, demonstrating diverse whole-body motion tracking, balance under external perturbations, and forceful interaction.
This work introduces HumanTracker, a preference-aligned metric trained on 12K motion pairs containing 24K motions that better predicts human preferences and reveals contact and stability failures that kinematic metrics often miss.
Dai-En Liu, Zekun Qi, Jiayu Zeng et al.· 0 citations
HuMiT is presented, a whole-body teleoperation system built on a minimal reference target that requires only a minimal reference target, consisting only of root height, root velocity, and sparse keypoint positions at the current frame, yet achieves competitive or superior tracking superiority compared to methods relying on more diverse reference states.
Humanoid robots have recently demonstrated promising capabilities in real-world ball sports. However, achieving professional motion styles while maintaining strong task performance remains challenging. In this work, we propose AdaPT, an Adaptive Motion Planning and Tracking framework that learns professional tennis serving and rally styles directly from broadcast videos. This hierarchical design is motivated by the key insight that the planner generates stylistic kinematic motions, while the tracker executes them with minimal interference with planning. Despite its effectiveness in simulation, a substantial sim-to-real gap emerges: tracking performance inevitably degrades on real robots, and this degradation is partially overlooked by autoregressive planning and further compounded by noisy perception. To address these issues, our adaptation mechanism improves tracking robustness by learning to track randomized execution speeds, while conditioning the planner on a learned motion-speed adapter to mitigate compounding errors. Real-world experiments on the Unitree G1 demonstrate the effectiveness of our adaptation mechanism in bridging the sim-to-real gap. We further deploy AdaPT policies on the full-size Dobot Atom humanoid robot (1.7m) and demonstrate in-the-wild serving without motion capture. Beyond these results, our real-world experiments reveal both algorithmic and engineering insights for future humanoid ball-sports systems. Videos and code are available on our \href{https://humanoidtennis.github.io/AdaPT/}{project website}.
Tao Huang, Ruofei Liu, Xuchen Tang et al.· 0 citations