Di WuROBOTICS / SYSTEMS

AUTOMATION / ROBOTICS / AUTONOMOUS SYSTEMS

Connecting perception
to action.

I am an undergraduate student in Automation interested in robotics and autonomous systems. My work connects visual perception, motion control, embedded development and embodied policy experiments to interactive, verifiable systems.

Explore projectsAbout me

Interfaces, implementation, and verification in complete systems.

PERCEPTION → ACTION
  1. 01Perception
  2. 02Coordinate mapping
  3. 03Motion control
  4. 04Interaction
System schematic · not a device photo or execution record

01 / Selected work

Learning through systems.

Four projects across vision, control, embedded systems, and embodied policies.

01 / Robotics
SYSTEM / FLOW
  1. 01Camera images
  2. 02Three-class detection
  3. 03Calibration and plane mapping
  4. 04Commands and feedback decisions
  5. 05Robot motion
System schematic · not a device photo or execution record

Vision and Multimodal Four-Axis Robot

A desktop robot that recognizes objects, responds to commands and uses human feedback for cat-like trial and correction.

My role
Team member · Visual perception, multimodal interaction and system integration
  • YOLOv8
  • OpenCV
  • Vosk
  • MediaPipe

The historical training CSV records a best validation mAP@0.5 of 86.21%, a detection metric.

Read project
02 / Embedded
SYSTEM / FLOW
  1. 01Radar and visual sensing
  2. 02Visit and ring policy
  3. 03Browser WebRTC session
  4. 04Video and two-way audio
  5. 05Servo unlock demonstration and reset
System schematic · not a device photo or execution record

ESP32-P4/C5 Multimodal WebRTC Doorbell

Embedded audio and video, radar and visual presence sensing connected to browser calls and a servo door-lock demonstration.

My role
Independent development · Board adaptation and integration on Espressif examples
  • ESP-IDF
  • WebRTC
  • ESP32-P4/C5
  • LD2450

Historical logs record three LAN call-setup times of about 4, 6 and 4 seconds, compared with about 30 seconds initially.

Read project
03 / Learning & localization
SYSTEM / FLOW
  1. 01Three cameras and robot state
  2. 02Vision-language-action policy
  3. 03WebSocket policy service
  4. 0414D actions and simulation
  5. 05Task-level evaluation and failure analysis
System schematic · not a device photo or execution record

GOAI Bimanual Policy Training and Evaluation

Multicamera vision-language-action policy adaptation, training deployment, local closed-loop simulation and failure analysis.

My role
Individual competition project · Open-model adaptation, experiment organization and policy deployment
  • StarVLA
  • QwenOFT
  • WebSocket
  • Isaac Sim

The stable baseline achieved 33/600 and the 20k fine-tune 28/600 across 24 tasks each: 1,200 episodes in total.

Read project
04 / Control
SYSTEM / FLOW
  1. 01Targets and waypoints
  2. 02Forward and inverse kinematics
  3. 03Four trajectory methods
  4. 04PID tracking simulation
  5. 05Error and torque analysis
  6. 06URDF / PyBullet visualization
System schematic · not a device photo or execution record

Robot Trajectory Planning and Control Simulation

From a planar 2-DOF arm to a spatial 3-DOF model: comparing trajectories, PID tracking and multibody simulation.

My role
Independent development · Kinematics, trajectory planning and control simulation
  • MATLAB
  • PID
  • URDF
  • PyBullet

Archived MATLAB plots, CSV files and analysis for four trajectory methods form a complete planning-to-control simulation workflow.

Read project
View all projects17 public repositories · algorithms, systems, and coursework

02 / Practice

Skills grounded in projects

Understanding methods through the tasks, interfaces, and systems that use them.

01

Perception & vision

Python / OpenCV / YOLOv8

Object detection, camera calibration, and planar coordinate mapping.

02

Kinematics & control

MATLAB / PID / PyBullet

Forward and inverse kinematics, trajectory planning, tracking, and simulation.

03

Embedded & communication

ESP-IDF / ESP32-P4/C5 / WebRTC

Board porting, sensor interfaces, and audio/video integration.

04

Policy adaptation & evaluation

StarVLA / QwenOFT

Multi-view and state interfaces, training workflows, and local simulation evaluation.

03 / About

Connecting algorithms
to working systems.

I am an undergraduate student in Automation interested in robotics and autonomous systems. I want to understand perception, planning, and control as parts of complete systems. I value clear interfaces, traceable experimental records, and an accurate distinction between physical-device and simulation results.

04 / Contact

Let’s talk robotics
and systems.

Code, project notes, and validation boundaries are available on GitHub.

Visit GitHub