发表机构
Institute for AI Industry Research (AIR), Tsinghua University; The University of Hong Kong (HKU); Institute for Interdisciplinary Information Sciences (IIIS), Tsinghua University; School of Aerospace Engineering, Tsinghua University; Peking University (PKU)（清华大学人工智能产业研究院; 香港大学; 清华大学交叉信息研究院; 清华大学航天航空学院; 北京大学）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
MSK-Bench提出包含22个全身肌肉骨骼运动任务的基准，评估5种控制范式，并联合考量任务成功、鲁棒性与生理指标，为肌肉驱动人形机器人控制提供标准化测试平台。
AI 中文摘要
肌肉骨骼（MSK）人形机器人为研究全身运动控制提供了生理学上的具身基础，但其高维肌肉驱动、延迟的激活动力学以及冗余的肌肉-肌腱结构使得学习比扭矩驱动的人形机器人控制困难得多。现有的MSK基准在步态、假肢、灵巧手或特定挑战赛道方面仍然分散，导致全身肌肉驱动控制未能在标准化任务、方法和指标下得到充分评估。我们引入了MSK-Bench，一个包含22个全身运动控制任务的基准，这些任务分为三个逐步具有挑战性的类别：姿势稳定、常见运动行为和接触丰富的环境交互。在统一的任务协议和鲁棒性扰动下，MSK-Bench评估了5种代表性控制范式，包括基于奖励的强化学习、智能体奖励调优、潜在动作强化学习、模仿先验控制以及基于模仿先验的残差适应。除了任务成功和奖励之外，MSK-Bench还进一步报告了鲁棒性分析和生理导向的诊断，包括激活成本、关节平滑度和EMG包络相似性。我们的实证研究表明，具身感知探索和结构化动作表示提高了高维肌肉空间中的任务覆盖率，模仿先验增强了参考兼容的稳定和运动，但在接触丰富的环境不匹配下会退化，而残差适应可以在固定参考失败时恢复成功行为。我们进一步发现，任务成功的提高并不一定意味着生理一致性的提高，这凸显了联合评估任务性能、鲁棒性和生理行为的重要性。MSK-Bench为全身肌肉驱动的人形机器人控制提供了一个任务-方法-指标测试平台。
英文摘要
Musculoskeletal (MSK) humanoids provide a physiologically grounded embodiment for studying full-body motor control, but their high-dimensional muscle actuation, delayed activation dynamics, and redundant muscle--tendon structures make learning substantially harder than torque-driven humanoid control. Existing MSK benchmarks remain fragmented across gait, prosthetics, dexterous hands, or challenge-specific tracks, leaving full-body muscle-actuated control insufficiently evaluated under standardized tasks, methods, and metrics. We introduce MSK-Bench, a benchmark of 22 full-body motor-control tasks organized into three progressively challenging categories: postural stabilization, common locomotor behaviors, and contact-rich environmental interaction. Under unified task protocols and robustness perturbations, MSK-Bench evaluates 5 representative control paradigms, including reward-based RL, agentic reward tuning, latent-action RL, imitation-prior control, and residual adaptation over imitation priors. Beyond task success and reward, MSK-Bench further reports robustness analysis and physiology-oriented diagnostics, including activation cost, joint smoothness, and EMG-envelope similarity. Our empirical study shows that embodiment-aware exploration and structured action representations improve task coverage in high-dimensional muscle spaces, imitation priors enhance reference-compatible stabilization and locomotion but degrade under contact-rich terrain mismatch, and residual adaptation can recover successful behaviors when fixed references fail. We further find that improved task success does not necessarily imply improved physiological agreement, highlighting the importance of evaluating task performance, robustness, and physiological behavior jointly. MSK-Bench provides a task--method--metric testbed for full-body muscle-actuated humanoid control.
CommentsProject page: https://zzongzheng0918.github.io/MSK-Bench/
