Assembly Lines
New AI Tool Could Make Assembly Faster and More Efficient
An AI-powered tool uses text prompts to assemble a variety of objects out of Lego bricks.
PITTSBURGH—Engineers at Carnegie Mellon University (CMU) have developed an AI-powered tool that uses text prompts to assemble objects faster and more efficiently.
BrickGPT takes a simple word, such as “guitar,” and creates a step-by-step guide for a person or a robot to build a physically stable model of that object. The tool currently uses Lego bricks to create items.
“This research paves the way toward generative manufacturing, which is when people can use a generative model to design everyday objects they can build themselves,” says Jun-Yan Zhu, Ph.D., an assistant professor at CMU’s Robotics Institute. “They can build a chair, a sofa or a toy.
“This is a new frontier, a new usage of these models beyond creating social media videos or photos,” explains Zhu. “These brick toy pieces are a simple medium, and it’s a starting point.”
Zhu believes this fusion of AI and robotics could speed up the process of designing and building many types of products.
“This could be a huge benefit to the manufacturing world,” claims Changliu Liu, Ph.D., an associate professor at the Robotics Institute. “It takes a long time to turn ideas into a physical design and prototype. But, if you can integrate generative AI into the process, it can significantly improve efficiency and reduce the roadblocks to kicking off projects.”
BrickGPT can currently assemble 21 types of items out of Lego bricks, including a birdhouse, piano and sofa. To generate a sofa, CMU engineers type the word “sofa” into BrickGPT, which generates a 3D model. Then, an algorithm transforms the model into brick structures and BrickGPT checks to ensure the structure is stable. A person or robotic arm can follow the steps and build the sofa.
Looking for quick answers on assembly and manufacturing topics? Try Ask ASM, our new smart AI search tool. Ask ASM
To train BrickGPT, the engineers generated StableText2Brick, a dataset containing more than 47,000 brick structures made from more than 28,000 unique 3D objects accompanied by detailed captions.
An autoregressive large language model predicts the next brick based on the previous one, ensuring that the structure is stable and won’t fall over. If there’s an error along the way, BrickGPT goes back and eliminates unstable points to guarantee the structure's stability.
Zhu and his colleagues hope to scale up this model, enabling it to generate more than the current 21 objects. They also plan to expand the diversity of their library pieces to increase the accuracy and complexity of generated designs.Looking for a reprint of this article?
From high-res PDFs to custom plaques, order your copy today!
