PROJECTS

Enhancing Design with Multi-Objective Generative Adversarial Networks

Master Thesis

2019.10 - 2020.10

Used BwUniCluster and Google Colab for training MOGAN. Codes were written in Python and Grasshopper. Applied libraries including Tensorflow, Numpy, Matplotlib, etc.
Conducted an in-depth study of design theories to propose a framework whereby a computer can holistically combines multiple classes of information in a generative system to generate design solutions using data-driven method.

Master Seminar Project

2020.01 - 2020.02

Trained a convolutional autoencoder using topographic data and satellite images obtained from the Google Maps API. And interfaced an XBOX 360 Kinect to scan topography data from a sandbox to infer potential landscape imagery.
Created a tactile design tool which a user may generate synthetic landscapes from a sand-crafted topography.

Master Seminar Project

2019.06 - 2019.07

Trained a customized object detection and tracking YOLO network using Tensorflow. Used OpenCV and trained network for computer vision and locating objects relative to the stereo camera.
Controlled the industrial robot by KRL code using serial communication.
Designed and 3D printed an end-effector for KUKA robot with Raspberry Pi, stereo camera and a customized gear system to accurately apply blocks of paint for censorship.

Master Design Studio

2018.10 - 2019.10

Digitally modeled natural fiber winding paths in Rhino and Grasshopper and provided digital supports for robotic fabrication and structure analysis teams.
Proposed various design iterations for the pavilion in Botanical garden of Freiburg. Collectively designed and crafted scaled model of final pavilion.

Master Seminar Project

2019.11 - 2020.02

Simulated the rolled and the flattened geometries in Rhinoand Grasshopper, and used Karamba and Galapagos to optimize the geometry and inform fiber layout.
Designed and crafted a stool made of natural fiber composite using an innovative cost-effective waste-free molding technique through rolling.

Rennteam Uni Stuttgart

2019.02 - 2019.04

Used Adobe Illustrator to design and create contents for the dashboard interface used in student formula racing car.
Developed the software for the dashboard interface using Qt. The front end is written in Javascript and the backend used C++ and CAN-Bus to interface vehicle data.

Master Seminar Project

2019.01 - 2019.02

Programmed agent behaviors of Physarum Polycephalum using Python in Rhino and Grasshopper environment to simulate its birth, death, and habituation behaviors. We extended its 2D behavior to 3D geometric generation.
Generated and 3D printed intricate geometries using agent-based modeling.

Master Seminar Project

2018.12 - 2019.02

Used Karamba and Galapagos in Grasshopper for multi-objective optimization to minimize the self-weight and displacement of the truss system.
Programmed the generation of a truss bridge through parametric modeling.

Research Project

2016.05 - 2018.05

Optimized grasshopper algorithms developed by Professor Lucia.
Explored and proposed potential research directions.
Produced exhibition contents by adapting various workflows using multiple programs.
Assisted on curating exhibitions.

Undergraduate Design Project

2017.01 - 2017.05

Applied biomimetic thinking by researching the skin of Cuttlefish, which changes color and texture to protect itself.
Developed a membrane that changes its color and texture as an ever-changing model taking light as an input and adapt itself accordingly.
Designed and produced components controlled by Arduino.
Used Ladybug component in grasshopper for sunlight analysis for geometry prediction.

Undergraduate Design Project

2016.01 - 2016.05

Used Processing to generate images with a sense of depth, and constructing models to capture that perceptional space.
Designed and crafted models, one of which used arches floating in the space as frames for winded strings.