利用外骨骼式辅助装置实现物体操控及日常任务执行的多模态、连续且安全的学习

本研究重点在于利用机器人或臂/手外骨骼来编程及半自动执行日常任务，以协助上肢运动功能严重受损的人士。该项目旨在开发能够适应用户残余运动能力的直观机器人辅助系统，从而提升其在日常活动中的参与度。申请者应具备工程学、计算机视觉及人工智能方面的背景，尤其是强化学习领域的相关知识。

Job description: Keywords robot learning, assistive robotics Project description: This thesis aims to study the programming and semi-automatic execution of everyday tasks by a robot or an arm/hand exoskeleton to assist movement. Based on state-of-the-art tools for humanoid robotic manipulation, the challenge will be to pilot robotic assistance in a natural or at least intuitive, acceptable and trustworthy manner for the performance of daily tasks by a Person with a Disability (PSH) with a significant loss of mobility of the upper limb. Beyond the learning techniques used in humanoid robotics, the use of the exoskeleton will facilitate demonstrations and the learning of everyday tasks, particularly by an able-bodied person. Furthermore, during use, the user's residual movements can be interpreted as initiating signals and guide the intention, while allowing the execution of the task to be adapted persistently, and thus improve appropriation. References [1]: A. L. Benabid & al., “An exoskeleton controlled by an epidural wireless brain–machine interface in a tetraplegic patient: a proof-of-concept demonstration”, The Lancet Neurology, Volume 18, Issue 12, 2019, Pages 1112-1122, ISSN 1474-4422, ［会员资料链接］. [2]: C. Dune, A. Remazeilles, E. Marchand & C. Leroux, “Vision-based grasping of unknown objects to improve disabled people autonomy”, in Robotics: Science and Systems Manipulation Workshop: Intelligence in Human Environments, Zurich, Switzerland, 2008. [3]: C. Rolinat, M. Grossard, S. Aloui, & al. “Grasp space exploration method for an underactuated gripper using human initiated primitive grasps”, Int J Intell Robot Appl 7, 31–47 (2023). ［会员资料链接］ [4]: G. Raiola, X. Lamy and F. Stulp, 'Co-manipulation with multiple probabilistic virtual guides,' 2015 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), Hamburg, Germany, 2015, pp. 7-13, //doi:10.1109/IROS.2015.7353107 [5]: A. Orhan, D. Hoàng, O. Bruneau, B. Berret and F. Geffard, 'Using Artificial Demonstrations Emulating Human Movement Variability for a Learning-Based Exoskeleton Flow Controller,' in IEEE Robotics and Automation Letters, vol. 10, no. 9, pp. 9430-9437, Sept. 2025, doi:10.1109/LRA.2025.3592058 [6]: M. Sundermeyer, A. Mousavian, R. Triebel, and D. Fox, “Contact-GraspNet: Efficient 6-DoF Grasp Generation in Cluttered Scenes,” in 2021 IEEE International Conference on Robotics and Automation (ICRA), 2021, pp. 13438-13444. [7]: A. Murali & al., “GraspGen: A Diffusion-based Framework for 6-DOF Grasping with On Generator Training”, 2025, ［会员资料链接］ [8]: P. Wang & al., “DemoGrasp: Few-Shot Learning for Robotic Grasping with Human Demonstration”, 2021, ［会员资料链接］ [9]: C. Zheng & al., “Embracing Bulky Objects with Humanoid Robots: Whole-Body Manipulation with Reinforcement Learning”, 2025, ［会员资料链接］ [10]: Genesis AI Team, “The Role of Simulation in Scalable Robotics, Genesis World 1.0, and the Path Forward”, 2026, ［会员资料链接］ [12]: J. Luo, C. Xu, J.Wu, S. Levine, “Precise and Dexterous Robotic Manipulation via Human-in-the-Loop Reinforcement Learning”, 2025, ［会员资料链接］ [13]: Welte E and Rayyes R, “Interactive imitation learning for dexterous robotic manipulation: challenges and perspectives—a survey”, 2025 Front. Robot. AI 12:1682437. doi: 10.3389/frobt.2025.1682437 [14]: Y. Zhu, Z.Wang, J., Merel, A., Rusu, T. Erez, S. Cabi, , ... & N. Heess, “Reinforcement and imitation learning for diverse visuomotor skills”, 2018, ［会员资料链接］ [15]: Y. Kuang, H. Geng, A. Elhafsi, T. D. Do, P. Abbeel, J. Malik, ... & Y. Wang,. “Skillblender: Towards versatile humanoid whole-body loco-manipulation via skill blending”, 2025, ［会员资料链接］ [16]: M. Tian, Y. Li, S. Liu, A. Ijspeert, S. Calinon, “Spline Policy: A Structured Representation for Robot Policies”, 2026, ［会员资料链接］ [17]: Z. Sun, B. Pang, X. Yuan, X. Xu, Y. Song, R. Song, and Y. Li, “Hierarchical reinforcement learning with curriculum demonstrations and goal-guided policies for sequential robotic manipulation”, Engineering Applications of Artificial Intelligence, vol. 153, p. 110866, 2025, doi: 10.1016/j.engappai.2025.110866 [18]: S. Luo, M. Jiang, S. Zhang, & al. “Experiment-free exoskeleton assistance via learning in simulation”, Nature 630, 353–359 (2024). ［会员资料链接］ [19]: D. P. Losey, C. G. McDonald, E. Battaglia, M. K. O'Malley, 'A Review of Intent Detection, Arbitration, and Communication Aspects of Shared Control for Physical Human–Robot Interaction.', 2018, ASME. Appl. Mech. Rev. January 2018; 70(1): 010804. ［会员资料链接］ [20] Duminy, N., Nguyen, S. M., Zhu, J., Duhaut, D., and Kerdreux, J. (2021). Intrinsically Motivated Open-Ended Multi-Task Learning Using Transfer Learning to Discover Task Hierarchy. Applied Sciences, 11(3). Mdpi. [21]Ma, Z., Nguyen, S.M., and Xu, P. (2025). Can LLMs Translate Human Instructions into a Reinforcement Learning Agent's Internal Emergent Symbolic Representation? NOVEL workshop - ITL4HRI. [22] Retzlaff, C. O., Das, S., Wayllace, C., Mousavi, P., Afshari, M., Yang, T., Saranti, A., Angerschmid, A., Taylor, M. E., and Holzinger, A. (2024). Human-in-the-Loop Reinforcement Learning: A Survey and Position on Requirements, Challenges, and Opportunities. Journal of Artificial Intelligence Research, 79(359--415). AI Access Foundation. [23] Baraka, K., Idrees, I., Kessler Faulkner, T., Biyik, E., Booth, S., Chetouani, M., Grollman, D. H., Saran, A., Senft, E., Tulli, S., Vollmer, A.-L., Andriella, A., Beierling, H., Horter, T., Kober, J., Sheidlower, I., Taylor, M. E., van Waveren, S., and Xiao, X. (2025). Human-Interactive Robot Learning: Definition, Challenges, and Recommendations. ACM Transactions on Human-Robot Interaction, . Association for Computing Machinery (ACM)

所需个人背景与技能要求： 来自知名工程院校的硕士或硕士2级学生，需具备良好的信息技术工具与计算机视觉技能，并拥有扎实的人工智能基础（理想情况下为强化学习相关领域）。申请人还必须对机器人技术有浓厚兴趣，乐于从事实验性工作（参与个人机器人项目将更为理想），且对健康与残疾问题具有敏感度。