Huaze Liu 刘铧泽

Hi, I'm Huaze, also go by Patrick. I'm a visiting student at Georgia Institute of Technology working on humanoid whole-body loco-manipulation, agentic self-improvement for whole-body control, and tactile sensing, advised by Prof. Ye Zhao.

Previously, I graduated from Harvey Mudd College with high distinction, where I worked with Prof. Adyasha Mohanty on sensor fusion for trustworthy autonomous navigation. I also worked on humanoid robot control and dexterous manipulation under the supervision of Prof. Michael Tolley and Prof. Hao Su at UC San Diego.

I'm interested in how tactile feedback can inform a robot's high-level planning to enable whole-body tasks that involve frequent contact or depend on precise contact control.

Email  /  Github  /  Linkedin  /  Google Scholar

My research focuses on contact-aware whole-body loco-manipulation for humanoid robots. Beyond sensing at the hands, I explore how tactile information from the arms, torso, and legs can reveal the robot's physical interactions with its surroundings. I investigate how to integrate these signals into high-level frameworks, including AI agents and vision-language-action (VLA) models, so that task plans account for where and how the body makes contact. I also explore agentic self-improvement through tactile feedback, connecting high-level planning with whole-body control for long-horizon tasks in which contact is essential.

Beyond robotics, I'm also interested in AI agents, computer vision, and ML theory. I regularly maintain and build new projects here.

Built a manual-grounded AI support agent for the Vulcan OmniPro 220 welder using retrieval over bundled PDF manuals and the Claude Agent SDK, returning page-level citations, manual visuals, and structured wiring/troubleshooting helpers through a Next.js chat UI.

A reproducible workflow for removing dynamic objects from videos by integrating YOLO for detection, OSTrack for tracking, SAM2 for segmentation, and DiffuEraser for diffusion-based inpainting. The system ensures temporal consistency and high-quality background restoration.

Developed a dual-robot system for autonomous emergency response, where an "explorer" robot uses vision-based detection and an EKF/UKF to locate targets and a "rescuer" robot uses the A* algorithm to navigate to them. The framework is validated in the Webots simulator.

Fine-tuned DETR (End-to-End Object Detection with Transformers) for traffic object detection task using KITTI object detection dataset, achieving 71.1% mAP with 18.1ms latency.

Activities: Member of Harvey Mudd RoboSub Team (2022-2023)

Awards: Dean's List (2022-2025), $11,000 Shanahan Projects Fund (2024), $3,000 Leeds Student Travel Grant (2025)