Machine understanding artificial intelligence

At the Future Work Summit held on January 12, 2022, listen to the opinions of CIOs, CTOs, and other C-level and senior managers on data and AI strategies. to know more information

(Reuters)-In a factory south of Toyota City, Japan, robots began to share the work of quality control inspectors, as the pandemic accelerated the transformation of Toyota’s boasting "Go and See" system, which helped in the 20th century Completely change the century of mass production. In Musashi Seimitsu's auto parts factory, a robotic arm picks up and rotates a bevel gear and scans its teeth under light to find surface defects. The inspection takes about two seconds—similar to a well-trained employee who inspects about 1,000 units per shift.

"Checking 1,000 exactly the same things every day requires a lot of skills and expertise, but it's not very creative," CEO Hiroshi Otsuka told Reuters. "We want to free workers from these tasks."

For a long time, global manufacturers have been using robots in production, leaving the tricky work of finding defects to humans. However, social distancing measures taken to prevent the spread of COVID-19 have prompted people to reconsider the factory floor. This has prompted increased use of robotics and other technologies for quality control, including remote monitoring that had been used before the pandemic.

In Japan, this method represents a serious departure from the "genchi genbutsu" observation methodology developed as part of the Toyota production system, which Japanese manufacturers have almost embraced with almost religious enthusiasm for decades. This process requires workers to constantly monitor all aspects of the production line to find abnormalities, and makes quality control one of the last manpower controls in other automated factories.

However, even at Toyota itself, when asked to automate more on-site procedures, a spokesperson said: "We are always looking for ways to improve the manufacturing process, including automating the process where it makes sense."

Improvements in artificial intelligence (AI) have been accompanied by cheaper equipment and stricter quality requirements from customers.

"We are increasingly discovering that there is a gap between the quality of products produced on conventional production lines and the quality required by customers," said Kazutaka Nagaoka, chief manufacturing officer of Apple supplier Japan Display and many automakers.

Nagaoka said: "The quality of products produced on automated production lines is much higher and more consistent."

However, automated inspections are challenging because robots need to be taught to recognize the tens of thousands of possible defects in a particular product and apply the learning immediately. Musashi Seimitsu's low defect rate per 50,000 products prevents the company from having enough defect examples to develop efficient AI algorithms. But the solution came from Israeli entrepreneur Ran Poliakine, who applied the artificial intelligence and optical technology he used in medical diagnosis to the production line. His idea is to teach the machine to find the good, not the bad, by building the algorithm on as many as 100 perfect or near-perfect units—modifying the so-called golden sample.

"If you observe human tissue, you are teaching algorithms what is good and what is bad, and you only have one second to make a diagnosis," he said.

After the breakthrough, Poliakine's startups SixAI and Musashi Seimitsu established MusashiAI, a joint venture that develops and employs quality control robots, which is the first in this field.

Poliakine said that since the new coronavirus spread to the world in March, the number of inquiries from automakers, component suppliers and other companies in Japan, India, the United States and Europe has quadrupled.

"COVID-19 has accelerated this process. Now everything is on steroids because working from home shows that remote work can work," he said.

Earlier this year, Magneti Marelli, an auto parts manufacturer with operating headquarters in Japan and Italy, also started using artificial intelligence quality inspection robots in a factory in Japan. The company told Reuters last month that it hopes artificial intelligence will be used in quality inspection. Play a greater role in the In the next few years.

Printer manufacturer Ricoh plans to automate all production processes of drum units and toner cartridges at one of its Japanese factories by March 2023. The robot has performed most of the process, and since April, technicians have been monitoring equipment on the factory floor at home.

Kazuhiro Kanno, general manager of Ricoh's printer manufacturing department, said: "Of course, when a problem occurs, you need to evaluate and implement the solution on site, but identification and confirmation are tasks that we can now complete at home."

Musashi Seimitsu will not disclose when it envisages fully automated factory floors, but Otsuka said artificial intelligence will complement rather than threaten the visitor system.

"Artificial intelligence doesn't ask'why? Why?' but humans do. We want them to make time to ask why and how defects occur," he said. "This will enable more people to find ways to continuously improve production. This is the purpose of'genchi genbutsu'."

(Reporting by Naomi Tajitsu and Makiko Yamazaki, additional reporting by Maki Shiraki and Noriyuki Hirata. Editing by David Dolan and Christopher Cushing.)

Please join Trinet on December 1st because they provide tips on how to avoid the pitfalls of excessive growth, how to return attention to consumers, and identify signs of difficulty in your team.

Hear the opinions of CIOs, CTOs, and other C-level executives on data and AI strategies.

© 2021 VentureBeat. all rights reserved.

We may collect cookies and other personal information from your interactions with our website. For more information about the categories of personal information we collect and the purposes for which we use them, please review our collection notice.