Research
Research Outputs
AI-based mobility/product design (Generative design)
What if artificial intelligence can design engineering systems on its own?
Smart Design Lab (SDL) wants to open a new paradigm in the manufacturing industry with AI-based engineering design. It combines physics-based dynamics with data-based deep learning technologies, and builds a virtual product development platform through digital transformation. This innovatively reduces the time and cost of product development and enables design to succeed in the marketplace.
AI-based generative design, which combines artificial intelligence and engineering design technologies (optimal design, CAD/CAM/CAE, HCI, etc.), allows artificial intelligence to achieve designs that satisfy engineering performance, aesthetics, and economy on its own. Develop and integrate the elemental technologies required for the four steps below.
- ① Design Generation: Technology that mass-creates aesthetic and engineering valid designs based on historical design data
- ② Design Evaluation: Technology to predict the various engineering performance, manufactureability, and originality of the generated design in real time and explain the cause
- ③ Design Optimization: Technology to derive optimal design proposals in real time to meet target performance and constraints
- ④ Design Recommendation: Technology that predicts customer preferences for design proposals in real time and recommends marketable designs
![SDL LAB 1 image3](https://mo.kaist.ac.kr/wp-content/uploads/2022/09/image3-1024x480.png)
![SDL LAB 2 image14](https://mo.kaist.ac.kr/wp-content/uploads/2022/09/image14-1024x467.png)
① Design Generation AI
![SDL LAB 3 image7](https://mo.kaist.ac.kr/wp-content/uploads/2022/09/image7.png)
![SDL LAB 4 2d 1](https://mo.kaist.ac.kr/wp-content/uploads/2022/09/휠2d-1.gif)
![SDL LAB 5 image11 1](https://mo.kaist.ac.kr/wp-content/uploads/2022/09/image11-1-1024x444.png)
![SDL LAB 6 3d 1](https://mo.kaist.ac.kr/wp-content/uploads/2022/09/휠3d-1.gif)
② Design Evaluation AI
![SDL LAB 7 image6 1](https://mo.kaist.ac.kr/wp-content/uploads/2022/09/image6-1-1024x465.png)
![SDL LAB 8 image10 1](https://mo.kaist.ac.kr/wp-content/uploads/2022/09/image10-1-1024x511.png)
![SDL LAB 9 image5](https://mo.kaist.ac.kr/wp-content/uploads/2022/09/image5-1024x546.png)
③ Design Optimization AI
![SDL LAB 10 image4](https://mo.kaist.ac.kr/wp-content/uploads/2022/09/image4-1024x335.png)
![SDL LAB 11 image9](https://mo.kaist.ac.kr/wp-content/uploads/2022/09/image9-1024x461.png)
Deep Learning Based Design Parameters Inverse Design
④ Design Recommendation AI
![SDL LAB 12 image17](https://mo.kaist.ac.kr/wp-content/uploads/2022/09/image17-1024x607.png)
![SDL LAB 13 image12](https://mo.kaist.ac.kr/wp-content/uploads/2022/09/image12-1024x589.png)