Wei-Teng (Wayne) CHU

📣 Wayne is actively seeking internship opportunities for Summer 2026! 📣

🌟 Highlight 🌟

Wayne dedicated much of his time to robotics and computer vision through projects, internships, and research. In the summer of 2023, he developed a project on dynamic obstacle avoidance using TurtleBot4 at UC San Diego, relying solely on depth camera inputs. His internship at ITRI further deepened his understanding of the challenges in bridging the Sim2Real gap, particularly in designing perception pipelines that remain robust in real-world conditions.

He conducted remote research in computer vision and robotics under the supervision of Prof. Weiming Zhi and Dr. Tianyi Zhang at the DROP Lab, CMU before beginning his studies at Stanford.

Education

Stanford University, CA

M.S. Student, Electrical Engineering

Concentration: Robotics / Computer Vision

• GPA: N/A
• Duration: Sep. 2025 - Apr. 2027 (Expected)

National Tsing Hua University, Taiwan

Bachelor of Science in Interdisciplinary Program of Engineering

Concentration: Electrical Engineering / Power Mechanical Engineering

• Graduated with Distinction
• GPA: 4.15 / 4.30
• Duration: Sep. 2020 - Jan. 2024

Work Experience

Stanford Artificial Intelligence Laboratory (SAIL)

Researcher @ Stanford Vision and Learning Lab

Sep. 2025 - In Progress

Current Work:
1. Constructed real-to-sim-to-real pipeline to generate “controllable digital cousins” to close sim-to-real gap in robot learning policy
2. Collected data in reality and developed real-to-sim replay method

Industrial Technology Research Institute (ITRI)

Automotive AI Algorithm Development Intern

Sep. 2024 - Nov. 2024, Mar. 2025 - Jul. 2025

Mask2Former Semantic Segmentation

We combined and re-annotated Mapillary and ADE20k datasets to fulfill requirements for quadruped robots and autonomous vehicles. It was then converted to COCO format for the training of custom Mask2Former model using NVIDIA TAO Toolkit. Lastly, the re-trained model was integrated into ROS2 for real-time video inference on robotic and vehicle platforms at about 15 FPS.

Person Re-Identification (Re-ID)

We built a real-time pedestrian detection and re-ID pipeline using YOLO11n and OSNet. The pipeline employed cosine similarity for feature matching and it exported models in .onnx and TensorRT .engine formats for model acceleration. We successfully achieved real-time inference at about 25 FPS for robotics and automotive use cases.

Sim2Real Quadruped Robot Terrain Traversal

We developed an elevation mapping workflow for quadruped robots Unitree Go2 by integrating Gazebo and point clouds from Intel RealSense depth camera. Moreover, a point cloud sampling & processing pipeline is integrated with reinforcement learning gait models in simulation. The most challenging part is to employed visual inertial odometry (VIO) to resolve sensor interruptions and implemented forward kinematics to compute foot coordinates of Unitree Go2. Finally, the system was deployed on Unitree Go2 for Sim2Real validation.

Foxconn

Technology Innovation Group of Chairman Office Intern

Jun. 2024 - Sep. 2024

We designed an intuitive Android car app to control A/C temperature of the vehicle via voice. The pipeline was composed of Azure Speech Service → Dialogflow intent parsing → CarAPI to operate the HVAC system.

[video]

Selected Research & Projects

Efficient Construction of Implicit Surface Models From a Single Image for Motion Generation

Mar. 2025 - Sep. 2025

[DROP Lab], Carnegie Mellon University

Advised by Dr. Tianyi Zhang, Prof. Matthew Johnson-Roberson, and Prof. Weiming Zhi
We present Fast Image-to-Neural Surface (FINS), a lightweight framework capable of reconstructing high fidelity signed distance fields (SDF) from a single image in 10 seconds. Our method fuses a multi-resolution hash grid and efficient optimization to achieve state-of-the-art accuracy while being an order of magnitude faster than existing methods. The scalability and real-world usability of FINS were also tested through robotic surface-following experiments, showing its utility in a wide range of tasks and datasets.

[arXiv]

Smart Fridge

Jun. 2024 - Sep. 2024

Google Hardware Product Sprint (HPS) Program

Developed a smart refrigerator with food pattern recognition and expiration tracking using Google Gemini. Built a recipe suggestion system based on soon-to-expire ingredients.

[poster] [slides] [code]

Laser-Assisted Guidance Landing Technology for Drones

Jan. 2023 - Nov. 2023

[HSCC Lab], National Tsing Hua University

Advised by Jang-Ping Sheu.
We propose a laser-assisted guidance approach to improve the landing accuracy of a drone where GPS creates a multi-meter error. This system merges embedded electronics, 3D-printed mechanical design, and low-power laser sensing to achieve 30-40 cm landing accuracy while also confirming the feasibility of laser-based localization for future unmanned aerial vehicle (UAV) applications that are either autonomous or remote.

[report] [slides] [video] [paper]

Dodging Dynamical Obstacles Using TurtleBot4 Camera Feed

Jun. 2023 - Aug. 2023

[MURO Lab], UC San Diego

Supervised by Jorge Cortés.
Implemented real-time dynamic obstacle avoidance for mobile robots using vision-based sensing and ROS2. The system employed the Turtlebot4 equipped with an OAK-D Pro camera to achieve precise obstacle tracking and smooth, collision-free navigation. It also integrated RRT* path planning and Bézier curve smoothing, demonstrating affordable and efficient vision-driven navigation in dynamic environments.

[report] [slides] [video] [paper]

Publications

(† Corresponding author)

• Efficient Construction of Implicit Surface Models From a Single Image for Motion Generation
  • Wei-Teng Chu, Tianyi Zhang, Matthew Johnson-Roberson, Weiming Zhi†
    arXiv [pdf]
    BibTeX

    @misc{chu2025efficientconstructionimplicitsurface,
      title={Efficient Construction of Implicit Surface Models From a Single Image for Motion Generation}, 
      author={Wei-Teng Chu and Tianyi Zhang and Matthew Johnson-Roberson and Weiming Zhi},
      year={2025},
      eprint={2509.20681},
      archivePrefix={arXiv},
      primaryClass={cs.RO},
      url={https://arxiv.org/abs/2509.20681}, 
    }
    

    Plain Text

    W.-T. Chu, T. Zhang, M. Johnson-Roberson, and W. Zhi, “Efficient Construction of Implicit Surface Models From a Single Image for Motion Generation,” arXiv preprint arXiv:2509.20681, 2025. [Online]. Available: https://doi.org/10.48550/arXiv.2509.20681
    

• Laser-Assisted Guidance Landing Technology for Drones
  • Yung-Ching Kuo, Wei-Teng Chu†, Yu-Po Cho, Jang-Ping Sheu
    MASS 2024 [pdf]
    BibTeX

    @INPROCEEDINGS{10723664,
      author={Kuo, Yung-Ching and Chu, Wei-Teng and Cho, Yu-Po and Sheu, Jang-Ping},
      booktitle={2024 IEEE 21st International Conference on Mobile Ad-Hoc and Smart Systems (MASS)},
      title={Laser-Assisted Guidance Landing Technology for Drones},
      year={2024},
      pages={670-675},
      doi={10.1109/MASS62177.2024.00107}
    }
    

    Plain Text

    Y. -C. Kuo, W. -T. Chu, Y. -P. Cho and J. -P. Sheu, "Laser-Assisted Guidance Landing Technology for Drones," 2024 IEEE 21st International Conference on Mobile Ad-Hoc and Smart Systems (MASS), Seoul, Korea, Republic of, 2024, pp. 670-675, doi: 10.1109/MASS62177.2024.00107
    

• Dodging Dynamical Obstacles Using Turtlebot4 Camera Feed
  • Wei-Teng Chu†
    PMAE 2023 (The Best Oral Presentation Award) [pdf]
    BibTeX

    @InProceedings{10.1007/978-981-97-4806-8_17,
      author="Chu, Wei-Teng",
      editor="Mo, John P. T.",
      title="Dodging Dynamical Obstacles Using Turtlebot4 Camera Feed",
      booktitle="Proceedings of the 10th International Conference on Mechanical, Automotive and Materials Engineering",
      year="2024",
      publisher="Springer Nature Singapore",
      pages="195--204",
      isbn="978-981-97-4806-8"
    }
    

    Plain Text

    Chu, WT. (2024). Dodging Dynamical Obstacles Using Turtlebot4 Camera Feed. In: Mo, J.P.T. (eds) Proceedings of the 10th International Conference on Mechanical, Automotive and Materials Engineering. CMAME 2023. Lecture Notes in Mechanical Engineering. Springer, Singapore. https://doi.org/10.1007/978-981-97-4806-8_17
    

Selected Awards

• 2024 Member of the Phi Tau Phi Scholastic Honor Society (Top 1% of Graduands in NTHU)
• 2024 Outstanding Graduate Award of NTHU (Top 1 in the Department)
• 2024 The Excellence Award in the Undergraduate Research Competition
• 2024 The 2nd Place in Undergraduate Research Poster Interpretation Competition
• 2023, 2019 Dr. Chu Shun-I ZyXEL Scholarship (Top 15 in the Total of 2200 Seniors)
• 2023 UC San Diego J. Yang Scholarship Award
• 2023, 2022, 2021 NTHU & IPE Outgoing Exchange Student Scholarship
• 2022 Jason International Charity Fund Scholarship Award by Acer Inc.
• 2022 The Excellence Award in the Asia-Pacific Mechanics Contest for College Students
• 2021, 2020 Academic Achievement Award of NTHU