Yupeng Wang

CUHK MAE PhD Student

Yupeng Wang

I study human-like motion generation and muscle-driven control for super-limb collaboration, building on prior work in surgical robotics, AR-assisted intervention, and robot autonomy.

Department of Mechanical and Automation Engineering, The Chinese University of Hong Kong Supervisor: Prof. Darwin Lau

Email View Projects

Research Profile

Human-centered robotics across interaction, autonomy, and intervention.

My current PhD research studies how a supernumerary robotic limb can anticipate a user's near-future behavior and coordinate assistance in time, so the human and the extra limb can complete shared tasks naturally.

Human behavior prediction Intent-aware assistance Super-limb collaboration Muscle-driven control AR-assisted surgical robotics Surgical robot teleoperation

Current PhD Direction

Behavior prediction and assistive control for human-super-limb collaboration.

The public-facing goal of this work is to make a super-limb collaborate as a proactive partner rather than a passive tool. The system should infer what the user is likely to do next from movement, task context, and muscle-driven cues, then decide when and how the additional limb should help.

In a successful interaction, the human keeps agency over the task while the super-limb predicts the next action and completes complementary subtasks such as stabilizing an object, positioning a tool, or preparing the next manipulation step.

Research focus
Human behavior prediction, muscle-driven modeling, and intent-aware robot assistance
Research target
Predictable, timely, and user-aware collaboration between a person and a super-limb
Concept image of a person collaborating with a predictive super-limb assistant
Concept visualization of prediction-aware human-super-limb collaboration.

Selected Projects

Previous work in surgical robotics, augmented reality guidance, and robot learning.

Generalized AR-assisted surgical robot platform overview

CUHK Robotics Research

Generalized AR-assisted surgical robot platform

Developed a reusable AR interface for virtual-to-real calibration and robot teleoperation. The platform supports waypoint control, end-effector dragging, single and multiple targets, and multiple tool orientations.

Validated across bone drilling, tracheostomy planning, and ultrasound scanning scenarios.

Automatic calibration and dynamic registration workflow for deformable tissue

Advanced Intelligent Systems

Automatic virtual-to-real calibration and dynamic registration for ESD

Built an AR-assisted workflow for deformable tissue in endoscopic submucosal dissection. The system combines a SuperGlue-based 6D pose estimator, Metric3D depth filtering, PyBullet simulation, Unity, and HoloLens 2 visualization.

Ex vivo validation: 3.11 +/- 0.56 mm AR calibration error and 3.20 +/- 1.96 mm dynamic registration error.

Trajectory prediction examples for ESD circumferential incision

2025 IROS Workshop

Shared-control flexible endoscopic robot for semi-autonomous circumferential incision

Proposed a trajectory prediction and execution pipeline for ESD incision, integrating safety margins, geometric trajectory planning, AR gesture control, Unity GUI teleoperation, and automatic trajectory following.

Received Best Workshop Paper Award, Second Prize, at the 2025 IROS Workshop on Continuum Robots for Surgery.

AR-assisted incision guidance for tracheostomy planning

ROBIO 2024 Finalist

Multimodal AR-assisted incision guidance for tracheostomy planning

Integrated ultrasound-derived anatomical information with AR visualization to support cricothyrotomy and tracheostomy incision localization on phantom models.

Reported high localization precision with mean central positioning errors below 0.31 mm in evaluated scenes.

Navigational augmented reality workflow for robotic drilling

ARSO 2024

Navigational augmented reality for robotic drilling

Built a robot-assisted drilling workflow that combines AR navigation, computer-assisted orthopedic surgery planning, 3D Slicer registration, ROS, MoveIt, and UR5 execution.

Evaluated target placement and robotic drilling on porcine and phantom femur experiments.

UR manipulator drawing a sand painting pattern for a design studio project

Teaching Assistant

UR controller tutorials for MAE-Architecture robotic design

Supported MAE students and Architecture students in a collaborative robotics studio, helping teams translate personal design concepts into robot-executable tasks using a UR manipulator.

Students designed tasks such as sand painting, carving, and bubble blowing, then implemented them as robot-executed motions.

Publications

Selected papers and workshop contributions.

  1. Y. Wang, H. Gao, A. Wang, and H. Ren, "Automatic Virtual-to-real Calibration and Dynamic Registration of Deformable Tissue for Endoscopic Submucosal Dissection," Advanced Intelligent Systems.
  2. Y. Wang, H. Gao, X. Yang, X. Zuo, and H. Ren, "Shared-Control Flexible Endoscopic Robot Enabling Semi-Autonomous Circumferential Incision by Geometric Trajectory Planning with Safety Margins for Endoscopic Submucosal Dissection," 2025 IROS Workshop on Continuum Robots for Surgery.
  3. W. Yue, Y. Wang (co-first author), S. Chen, H. Gao, and H. Ren, "Multimodal Augmented Reality Assisted Incision Guidance for Preoperative Tracheostomy Planning," ROBIO 2024.
  4. H. Ho, Y. Wang (co-first author), A. Wang, L. Bai, and H. Ren, "Web-based Augmented Reality with Auto-Scaling and Real-Time Head Tracking towards Markerless Neurointerventional Preoperative Planning and Training of Head-mounted Robotic Needle Insertion," ROBIO 2024.
  5. Y. Wang, Y. Ma, and H. Ren, "Navigational Augmented Reality for Robotic Drilling," ARSO 2024.

Honors

Awarded for work on semi-autonomous ESD incision.

Best Workshop Paper Award certificate from 2025 IROS Workshop

Best Workshop Paper Award, Second Prize

2025 IROS Workshop on Continuum Robots for Surgery, for the work on shared-control flexible endoscopic robot-assisted circumferential incision.

Additional academic honors include CUHK M.Sc. Department Admission Scholarship, Dr. Raymond Leung Scholarship, Dean's List, and CUHK(SZ) Dean's List across undergraduate study.

Education

Training across mechanical automation, electronic engineering, and computer engineering.

Current

PhD Student
Mechanical and Automation Engineering

The Chinese University of Hong Kong
Supervisor: Prof. Darwin Lau

2023 - 2024

M.Sc.
Electronic Engineering

The Chinese University of Hong Kong
GPA: 3.75/4.00

2019 - 2023

B.Eng.
Computer Engineering

The Chinese University of Hong Kong, Shenzhen
Major GPA: 3.872/4.00
Rank: 3/101