MIT Startup Inkbit Is Using Artificial Intelligence to Improve 3D Printing

3D printing has been revolutionary for prototyping parts, because the technology allows engineers to evaluate and iterate their designs…

Cameron Coward
5 years ago3D Printing

3D printing has been revolutionary for prototyping parts, because the technology allows engineers to evaluate and iterate their designs quickly before shelling out the massive amount of money associated with creating production tooling for manufacturing. But 3D printing itself has largely remained unsuitable for actual production. There are three major reasons for that: 3D printing is relatively slow, relatively expensive per unit, and doesn’t match the quality of traditional manufacturing processes like injection molding. Now, a startup out of MIT called Inkbit wants to change that with computer vision and machine learning.

Even the most advanced 3D printers today have difficulty getting close to the surface finish quality that we have come to expect in consumer products. That’s due to the materials used and the nature of the printing process, which relies on laying down successive layers of material. Most 3D printers are open-loop systems, which means they don’t receive any feedback about their output and can’t correct for errors. The 3D printers being developed by Inkbit are closed-loop, and use computer vision combined with machine learning to monitor the print job as it’s happening and fix errors in real time. That dramatically improves the quality of prints, particularly with materials that are traditionally difficult to 3D print such as flexible rubber-like polymers.

Aside from improving print quality, Inkbit’s technology also allows for new innovations in 3D printing. For example, computer chips or sensors can be placed into an object as it’s being printed, effectively embedding the electronics in a way that hasn’t been practical before. Right now, Inkbit only has one 3D printer: a 16 print head inkjet model. That uses a custom high speed OCT (Optical Coherence Tomography) sensor, a technology usually used in ophthalmology, that can see resolutions much narrower than a human hair. In conjunction with a machine learning system, that can be used to correct microscopic errors in the print. For now, this printer will be evaluated by Johnson & Johnson, but Inkbit hopes to start selling their printers next year.

Cameron Coward
Writer for Hackster News. Proud husband and dog dad. Maker and serial hobbyist. Check out my YouTube channel: Serial Hobbyism
Latest articles
Sponsored articles
Related articles
Latest articles
Read more
Related articles