{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,8,26]],"date-time":"2026-08-26T09:19:32Z","timestamp":1787735972536,"version":"build-2784847793"},"reference-count":33,"publisher":"MDPI AG","issue":"9","license":[{"start":{"date-parts":[[2026,8,26]],"date-time":"2026-08-26T00:00:00Z","timestamp":1787702400000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"funder":[{"name":"2025 Shanghai Key Technologies Research and Development Program\u2014\u201cNew Energy\u201d","award":["No. 25DZ3001401"],"award-info":[{"award-number":["No. 25DZ3001401"]}]}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":["Biomimetics"],"abstract":"<jats:p>Tendon-cable transmission can reduce distal-limb inertia in full-size humanoids, but its elasticity, hysteresis, backlash, and multi-joint coupling introduce state-dependent joint-to-motor discrepancies. We present a hierarchical whole-body tracking framework for the 28-DoF Droid X3 that separates high-level motion learning from transmission compensation. A reference-residual policy is trained in simulation by single-stage proximal policy optimization (PPO) using a unified robot-space motion representation, globally anchored tracking rewards, hierarchical hard-example sampling, and tendon-oriented domain randomization. In simulation checkpoint evaluation, more than 90% of 12,674 tested reference motions are completed. Independently, a state-conditioned mapper is trained offline through a differentiable motor\u2013joint forward model identified from physical motor-excitation data and connected in series between the frozen policy and the low-level motor controller. Randomized repeated Mapping-OFF\/ON trials are conducted on two nominally identical Droid X3 units. Within every robot\u2013motion block, the frozen PPO checkpoint, reference trajectory, controller settings, safety bounds, and frozen mapper weights are held fixed; complete trials are the statistical units. OFF converts desired joint positions with the robot-specific fixed static calibration, whereas ON feeds the complete policy-level desired-joint vector and measured plant state to the frozen mapper, which directly outputs the complete motor-position command. Across the complete physical trials, the aggregate action-completion rate is 68% with Mapping OFF and 79% with Mapping ON, an increase of 11 percentage points. Representative walk, squat, and dance trajectories illustrate lower tracking errors under Mapping ON, while individual frames and selected temporal fragments are used only for visualization.<\/jats:p>","DOI":"10.3390\/biomimetics11090607","type":"journal-article","created":{"date-parts":[[2026,8,26]],"date-time":"2026-08-26T08:11:15Z","timestamp":1787731875000},"page":"607","update-policy":"https:\/\/doi.org\/10.3390\/mdpi_crossmark_policy","source":"Crossref","is-referenced-by-count":0,"title":["Hierarchical Whole-Body Control for Tendon-Cable-Driven Humanoids via Reference-Residual Policy and Offline-Learned Tendon Mapping"],"prefix":"10.3390","volume":"11","author":[{"given":"Wencong","family":"Gan","sequence":"first","affiliation":[{"name":"School of Optical-Electrical and Computer Engineering, University of Shanghai for Science and Technology, Shanghai 200093, China"},{"name":"Institute of Machine Intelligence, University of Shanghai for Science and Technology, Shanghai 200093, China"},{"name":"Shanghai DroidUp Co., Ltd., Shanghai 200433, China"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Jiehui","family":"Chen","sequence":"additional","affiliation":[{"name":"School of Optical-Electrical and Computer Engineering, University of Shanghai for Science and Technology, Shanghai 200093, China"},{"name":"Institute of Machine Intelligence, University of Shanghai for Science and Technology, Shanghai 200093, China"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Qingdu","family":"Li","sequence":"additional","affiliation":[{"name":"Institute of Machine Intelligence, University of Shanghai for Science and Technology, Shanghai 200093, China"},{"name":"Shanghai DroidUp Co., Ltd., Shanghai 200433, China"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Haiming","family":"Mou","sequence":"additional","affiliation":[{"name":"Shanghai DroidUp Co., Ltd., Shanghai 200433, China"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Jianwei","family":"Zhang","sequence":"additional","affiliation":[{"name":"Department of Informatics, University of Hamburg, 20146 Hamburg, Germany"}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"1968","published-online":{"date-parts":[[2026,8,26]]},"reference":[{"key":"ref_1","doi-asserted-by":"crossref","first-page":"143","DOI":"10.1145\/3197517.3201311","article-title":"DeepMimic: Example-guided deep reinforcement learning of physics-based character skills","volume":"37","author":"Peng","year":"2018","journal-title":"ACM Trans. Graph. (TOG)"},{"key":"ref_2","unstructured":"Luo, Z., Cao, J., Merel, J., Winkler, A., Huang, J., Kitani, K., and Xu, W. (2024, January 7\u201311). Universal humanoid motion representations for physics-based control. Proceedings of the International Conference on Learning Representations (ICLR), Vienna, Austria."},{"key":"ref_3","doi-asserted-by":"crossref","unstructured":"He, T., Xiao, W., Lin, T., Luo, Z., Xu, Z., Jiang, Z., Kautz, J., Liu, C., Shi, G., and Wang, X. (2025, January 19\u201323). HOVER: Versatile neural whole-body controller for humanoid robots. Proceedings of the 2025 IEEE International Conference on Robotics and Automation (ICRA), Atlanta, GA, USA.","DOI":"10.1109\/ICRA55743.2025.11128549"},{"key":"ref_4","unstructured":"Chen, Z., Ji, M., Cheng, X., Peng, X., Peng, X.B., and Wang, X. (2025). GMT: General motion tracking for humanoid whole-body control. arXiv."},{"key":"ref_5","unstructured":"Yang, C., Sun, Y., Ye, P., Chen, X., Yu, C., and Chen, T. (2025). Efficiently learning general motion tracking policy for high dynamic humanoid whole-body control. arXiv."},{"key":"ref_6","doi-asserted-by":"crossref","unstructured":"Luo, Z., Yuan, Y., Wang, T., Li, C., Chen, S., Casta\u00f1eda, F., Cao, Z.A., Li, J., Minor, D., and Ben, Q. (2025). SONIC: Supersizing motion tracking for natural humanoid whole-body control. arXiv.","DOI":"10.1126\/scirobotics.aed4592"},{"key":"ref_7","unstructured":"Yao, Y., Luo, C., Du, J., He, W., and Lu, J.G. (2025). GBC: Generalized behavior-cloning framework for whole-body humanoid imitation. arXiv."},{"key":"ref_8","doi-asserted-by":"crossref","unstructured":"Wang, Y., Yang, M., Zeng, W., Zhang, Y., Xu, X., Jiang, H., Ding, Z., and Lu, Z. (2025). From experts to a generalist: Toward general whole-body control for humanoid robots. arXiv.","DOI":"10.52202\/085713-4941"},{"key":"ref_9","doi-asserted-by":"crossref","unstructured":"He, T., Luo, Z., He, X., Xiao, W., Zhang, C., Zhang, W., Kitani, K., Liu, C., and Shi, G. (2024). OmniH2O: Universal and dexterous human-to-humanoid whole-body teleoperation and learning. arXiv.","DOI":"10.1109\/IROS58592.2024.10801984"},{"key":"ref_10","unstructured":"Ze, Y., Chen, Z., Ara\u00fajo, J.P., Cao, Z.A., Peng, X.B., Wu, J., and Liu, C.K. (2025). TWIST: Teleoperated whole-body imitation system. arXiv."},{"key":"ref_11","unstructured":"Li, Y., Lin, Y., Cui, J., Liu, T., Liang, W., Zhu, Y., and Huang, S. (2025). CLONE: Closed-loop whole-body humanoid teleoperation for long-horizon tasks. arXiv."},{"key":"ref_12","unstructured":"Truong, T.E., Liao, Q., Huang, X., Tevet, G., Liu, C.K., and Sreenath, K. (2025). BeyondMimic: From motion tracking to versatile humanoid control via guided diffusion. arXiv."},{"key":"ref_13","unstructured":"Sun, Z., Huang, B.S., Peng, Y., Li, X., Ma, J., Sun, Y., Li, Z., Jiang, H., Gao, B., and Bing, Z. (2026). MOSAIC: Bridging the sim-to-real gap in generalist humanoid motion tracking and teleoperation with rapid residual adaptation. arXiv."},{"key":"ref_14","unstructured":"Zhao, S., Ze, Y., Wang, Y., Liu, C.K., Abbeel, P., Shi, G., and Duan, R. (2025). ResMimic: From general motion tracking to humanoid whole-body loco-manipulation via residual learning. arXiv."},{"key":"ref_15","unstructured":"Ara\u00fajo, J.P., Ze, Y., Xu, P., Wu, J., and Liu, C.K. (2025). Retargeting matters: General motion retargeting for humanoid motion tracking. arXiv."},{"key":"ref_16","unstructured":"Pan, Y., Qiao, R., Chen, L., Chitta, K., Pan, L., Mai, H., Bu, Q., Zhao, H., Zheng, C., and Luo, P. (2025). Agility meets stability: Versatile humanoid control with heterogeneous data. arXiv."},{"key":"ref_17","unstructured":"Xue, Y., Lin, Y., Dong, W., Tang, Y., Wang, J., Pang, J., Zhou, M., Liu, M., and Zhang, W. (2026). Scalable and general whole-body control for cross-humanoid locomotion. arXiv."},{"key":"ref_18","doi-asserted-by":"crossref","unstructured":"Zhang, Z., Wen, K., Xu, M., He, J., Li, C., Miki, T., Schwarke, C., Zhang, C., Peng, X.B., and Hutter, M. (2026). Learning whole-body humanoid locomotion via motion generation and motion tracking. arXiv.","DOI":"10.1109\/LRA.2026.3710365"},{"key":"ref_19","unstructured":"Li, J., Tang, B., and Wu, F. (2026). TeleGate: Whole-body humanoid teleoperation via gated expert selection with motion prior. arXiv."},{"key":"ref_20","unstructured":"Zhu, T., Cai, G., Yang, Z., Ren, G., Xie, H., Wang, Z., Wu, J., Wang, J., Yang, X., and Mu, Y. (2026). CLOT: Closed-loop global motion tracking for whole-body humanoid teleoperation. arXiv."},{"key":"ref_21","unstructured":"Li, Y., Ma, L., Lin, Y., Du, Y., Liu, M., Hu, K., Cui, J., Zhu, Y., Liang, W., and Jia, B. (2026). OmniClone: Engineering a robust, all-rounder whole-body humanoid teleoperation system. arXiv."},{"key":"ref_22","unstructured":"Wang, Y., Zhao, Q., Lau, Y.F., Yu, R., Tsui, H.W., Chen, Q., Wang, J., Pang, J., and Tan, P. (2026). HumanX: Toward agile and generalizable humanoid interaction skills from human videos. arXiv."},{"key":"ref_23","unstructured":"Zhang, Z., Lu, H., Lian, Y., Chen, Z., Liu, Y., Lin, C., Xue, H., Zeng, Z., Qi, Z., and Zheng, S. (2026). Learning athletic humanoid tennis skills from imperfect human motion data. arXiv."},{"key":"ref_24","unstructured":"Shanghai DroidUp Co., Ltd (2026, July 27). Bipedal Humanoid Robot Product Series. Available online: https:\/\/www.droidup.com\/product."},{"key":"ref_25","doi-asserted-by":"crossref","first-page":"144","DOI":"10.1145\/3450626.3459670","article-title":"AMP: Adversarial motion priors for stylized physics-based character control","volume":"40","author":"Peng","year":"2021","journal-title":"ACM Trans. Graph. (TOG)"},{"key":"ref_26","unstructured":"Lu, Q., Feng, Y., Shi, B., Piseno, M., Bao, Z., and Liu, C.K. (2025). GentleHumanoid: Learning upper-body compliance for contact-rich human and object interaction. arXiv."},{"key":"ref_27","unstructured":"Xiong, Z., Fang, L., Huang, J., Yamazaki, K., Zhang, H., and Gan, C. (2026). ExtremControl: Low-latency humanoid teleoperation with direct extremity control. arXiv."},{"key":"ref_28","unstructured":"Lee, K., Park, S., Park, G., Kim, M.J., and Park, J. (2026). Safety-critical whole-body control for humanoid robots via input-to-state safe control barrier functions. arXiv."},{"key":"ref_29","doi-asserted-by":"crossref","first-page":"228","DOI":"10.1038\/s41378-026-01322-0","article-title":"Fine-tuned multimodal large language model for autonomous state cognition system of shape-recognition 6-bar tensegrity integrated with flexible sensors","volume":"12","author":"Mao","year":"2026","journal-title":"Microsyst. Nanoeng."},{"key":"ref_30","doi-asserted-by":"crossref","unstructured":"Todorov, E., Erez, T., and Tassa, Y. (2012, January 7\u201312). MuJoCo: A physics engine for model-based control. Proceedings of the 2012 IEEE\/RSJ International Conference on Intelligent Robots and Systems (IROS), Vilamoura, Portugal.","DOI":"10.1109\/IROS.2012.6386109"},{"key":"ref_31","unstructured":"Mahmood, N., Ghorbani, N., Troje, N.F., Pons-Moll, G., and Black, M.J. (November, January 27). AMASS: Archive of motion capture as surface shapes. Proceedings of the IEEE\/CVF International Conference on Computer Vision (ICCV), Seoul, Republic of Korea."},{"key":"ref_32","unstructured":"Schulman, J., Wolski, F., Dhariwal, P., Radford, A., and Klimov, O. (2017). Proximal policy optimization algorithms. arXiv."},{"key":"ref_33","unstructured":"Schulman, J., Moritz, P., Levine, S., Jordan, M., and Abbeel, P. (2016, January 2\u20134). High-dimensional continuous control using generalized advantage estimation. Proceedings of the International Conference on Learning Representations (ICLR), San Juan, PR, USA."}],"container-title":["Biomimetics"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/www.mdpi.com\/2313-7673\/11\/9\/607\/pdf","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,8,26]],"date-time":"2026-08-26T08:40:04Z","timestamp":1787733604000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.mdpi.com\/2313-7673\/11\/9\/607"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,8,26]]},"references-count":33,"journal-issue":{"issue":"9","published-online":{"date-parts":[[2026,9]]}},"alternative-id":["biomimetics11090607"],"URL":"https:\/\/doi.org\/10.3390\/biomimetics11090607","relation":{},"ISSN":["2313-7673"],"issn-type":[{"value":"2313-7673","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026,8,26]]}}}