Jiahao Liu / 刘家灏

ROBOTICS · LEARNING · BUILDING

Jiahao Liu刘家灏

A robotics learner working toward full-stack capability

以全栈为目标的机器人小白

Jiahao Liu by a lake with Mount Fuji in the background.

I am currently a master's student at JSK Laboratory, The University of Tokyo. I graduated as an Outstanding Graduate from the School of Mechanical Science and Engineering at Huazhong University of Science and Technology. During my undergraduate studies, I was an exchange student at the Smart Robot Design Lab, Tohoku University.

我目前在东京大学 JSK 实验室攻读硕士。本科作为优秀毕业生毕业于华中科技大学机械学院；本科期间曾在日本东北大学 Smart Robot Design Lab 交换学习。

My current research combines impedance control and robot learning for contact-rich manipulation.

目前致力于结合阻抗控制与机器人学习，研究富接触任务中的机器人操作。

01 /

Vision 长期理想

I hope to help bring robots into the physical world and free people from repetitive and hazardous work. To do so, I want to understand the different layers of a robotic system—from mechanics and electronics to control, learning, and system integration—and eventually develop the ability to build complete robotic systems independently.

我希望机器人真正进入物理世界，帮助人类摆脱繁复、危险的工作。为此，我希望理解构成机器人系统的不同层级——从机械和电子，到控制、学习和系统集成——并最终具备独立构建完整机器人系统的能力。

02 /

Full-Stack Robotics 六维能力

Six connected dimensions. A direction for learning.
六个相互关联的维度，也是持续学习的方向。点击维度查看相关项目。

Full-stack robotics capability radar / 六维能力雷达Current self-assessment is shown with a solid line; the goal is shown with a dashed line. 当前能力以实线表示，目标以虚线表示。机械设计电路与嵌入式控制与优化机器学习数理理论系统集成
GoalCurrent
  1. Mechanical Design机械设计
  2. Electronics & Embedded电路与嵌入式
  3. Control & Optimization控制与优化
  4. Machine Learning机器学习
  5. Mathematical Foundations数理理论
  6. System Integration系统集成
03 /

Publications 论文

  1. 2026Submitted / 在投

    Continuous Manifold-Decomposed Impedance Retargeting for Contact-Rich Imitation Learning

    Jiahao Liu, Kento Kawaharazuka, Tasuku Makabe, Kei Okada

    I study how fixed-impedance teleoperation demonstrations can become continuous variable-impedance policies for learning. CMDIR introduces a continuous task-manifold impedance representation and Quality-to-Fast optimization to adjust contact behavior while preserving task response, then uses the optimized controllers to supervise imitation learning. I implemented demonstration retargeting, policy training, and robot execution, with experiments on planar wiping, curved wiping, and pushing.

    研究如何把固定阻抗遥操作示教转化为可用于学习的连续变阻抗控制策略。提出连续任务流形阻抗表示和 Quality-to-Fast 优化方法，在保持任务响应的同时调整接触行为，并用优化后的控制器监督模仿学习。完成从示教重定向、策略训练到机器人执行的实现，在平面擦拭、曲面擦拭和推动任务中开展实验。

    arXiv ↗YouTube ↗Project →
  2. 2026Submitted / 在投

    TRACT: Temporally Routed Action Chunks with Chronological Phase Authority for Contact-Rich Manipulation

    Jiahao Liu, Kento Kawaharazuka, Tasuku Makabe, Kei Okada

    I study phase transitions and stalled motion in action-chunk prediction for multi-stage contact-rich manipulation. TRACT routes predicted actions by task phase and uses a response-deficit integrator to compensate for motion suppressed by contact. I implemented training, inference, and hardware deployment and evaluated the method in multi-stage wiping.

    面向多阶段富接触操作，研究动作块预测中的阶段切换与接触受阻问题。提出按任务阶段路由动作的模仿学习策略，并结合响应不足积分器补偿接触中的运动受阻。完成模型训练、推理与实机部署，在多阶段擦拭任务中验证。

    arXiv ↗YouTube ↗Project →
  3. 2026Submitted / 在投

    MDIR: A Task-Manifold Impedance Retargeting Method for Contact-Rich Teleoperation

    Liu Jiahao, Kento Kawaharazuka, Tasuku Makabe, Kei Okada

    I developed task-manifold impedance retargeting to turn fixed-impedance demonstrations into variable-impedance controllers for contact tasks. The method separates work, exertion, and support requirements to moderate contact while preserving task response. I implemented the method and evaluated it on Franka Panda in planar wiping, pick-and-place, and pushing tasks.

    提出任务流形阻抗重定向方法，将固定阻抗示教转化为适合接触任务的变阻抗控制器。方法按工作、施力和支撑的不同需求组织控制，在保留任务响应的同时减缓接触冲击。完成方法实现及 Panda 实机实验，验证于平面擦拭、抓放和推动任务。

    arXiv ↗YouTube ↗Project →
  4. 2025Mechanical Systems and Signal Processing · 231, 112701 · published

    Evaluation and optimal design of suspended backpacks considering the system instability introduced by load fluctuations

    Tao Wang, Jiahao Wu, Jiahao Liu, Yang Liu, Jiejunyi Liang

    I studied how load fluctuations in suspended backpacks affect human walking stability and energy expenditure, contributing experimental data analysis and quantitative dynamics modeling. The work was published in Mechanical Systems and Signal Processing and informs backpack performance evaluation and design optimization.

    研究悬浮背包的负载波动如何影响人体步行稳定性与能耗，承担实验数据分析与动力学定量建模。成果发表于 Mechanical Systems and Signal Processing，为悬浮背包的性能评价与优化设计提供依据。

    DOI ↗Project →
04 /

Experience 经历

  1. Present / 在读

    The University of Tokyo

    东京大学

    Master's student / 硕士生JSK Laboratory

    Imitation Learning of Compliance Strategies for Contact-rich Tasks from Haptic Teleoperation Demonstrations

    研究课题：从力觉遥操作示范中模仿学习接触密集任务的柔顺策略。

    Excellent grades in all completed coursework.

    已修课程全优。

  2. 2025.09 — 2025.11

    桥介数物

    Intern / 实习生

    In a humanoid whole-body Mimic project, I was primarily responsible for data retargeting using joint trees and joint nearest-neighbor optimization.

    在人形机器人全身 Mimic 项目中，主要负责利用关节树和关节最近邻优化去进行数据重定向工作。

  3. 2023.10 — 2024.03

    Tohoku University

    东北大学

    Exchange student / 交换生Smart Robot Design Lab

    Worked on YOLOv8-based garment hem-point extraction for visual grasping, using a purpose-built dataset and a coupled two-model approach.

    围绕衣物视觉抓取开展研究，使用 YOLOv8、自建数据集与双模型结合方法提取衣物下摆特征点。

  4. 2020.09 — 2024.06

    Huazhong University of Science and Technology

    华中科技大学

    Mechanical Design, Manufacturing and Automation (Robotics) / 机械设计制造及其自动化（机器人）

    GPA: 3.96/4; 7 competition awards, 4 scholarships, and Outstanding Graduate.

    GPA 3.96/4，获 7 项竞赛奖励、4 项奖学金，并获优秀毕业生。