A system may watch a ‘live, breathing’ manufacturing line in real-time.
Elon Musk, the CEO of Tesla, often refers to the automotive factory as “the machine that builds the machine.” However, even the most highly automated facilities still employ a significant number of people.
They continue to play a crucial role in the highly complex process of making automobiles and trucks. Still, they must function just as well as their mechanical counterparts to keep the production line moving with quality and speed.
Sending quality control experts up and down the lines to look at the activity has always been the best way to iron out problems and ensure everything is going correctly. However, there is now a technique to automate that task with better outcomes than before.
Palo Alto-based Invisible AI was started by people who had worked in the self-driving car business for a long time. They saw that the artificial intelligence-based machine vision technology they were working on could be used in other ways and could be sold to the public long before self-driving cars became popular.
The company made a network of cams to watch a production line in real-time and find even the smallest problems.
“Productivity, safety, and quality are always at the top of our minds in manufacturing, especially in auto,” Eric Danzinger, CEO of Invisible AI, told Fox News Digital.
The self-contained units have binocular vision and processing onboard, making them easy to set up in a factory without having to connect to the factory’s networks.
“Our AI isn’t just about watching one workstation. It’s about seeing where you’re hitting production bottlenecks, where you’re seeing deviations from how the work is supposed to be done, and where you’re seeing problems like bad reaches that can cause physical problems for your workers,” Danzinger said. The cameras don’t need to be programmed with the assembly process. They only need to scan a single, correct cycle, and then the system can deduce the rest.
Danzinger said, “Our AI system looks at the raw pixels in the video to figure out how the work is being done. It then compares these patterns to see if someone follows a standard.” “An intelligent agent in the cams is doing all that so that a person doesn’t have to.
“If you put 100 cameras on one part of an assembly, you can see the line in 3D as it moves and changes.”
The price depends on the application, but Danzinger said it’s much cheaper than hiring a consulting team or trying to do the same work by hand, which isn’t possible given how much the system can do.
Since they are self-contained, all cams can be set up between shifts in a few days.
Danzinger said, “Our system has become the place you can go to help front-line workers understand what’s going on.”
“A million different things are going on. People are getting sick, sellers are sending bad parts, and tools are breaking down. To know what’s going on, what’s the most important thing to fix, and how to make my numbers. The most important thing is that.”
Invisible AI has 12 car parts providers and four original equipment manufacturers as clients. One of these is Toyota, which uses the system in an Indiana plant.
Toyota didn’t comment for this report, but when the project was announced last year, Senior Engineer Jihad Abdul-Rahim said, “Invisible AI is not only helping us find ways to improve the assembly lines, but we’re also always finding new ways to use their technology, like using ergonomics analysis to prevent injuries before they happen.”
Danzinger said that information about the company’s other users and how they use the system is private, so Invisible AI can’t discuss them.
Regarding privacy, the system doesn’t have facial recognition technology and can hide videos of people’s faces. But the point is to give direct feedback, so it is not a completely anonymous tool for analysis.
“Most of what we see is people being able to speak up and say, ‘This is broken. We need help fixing it,'” Danzinger said, “and actually getting an answer.”