Nissan improving paint quality with AI and machine learning
By onMarket Trends | Technology
Nissan has compiled a library of more than 500,000 vehicle paint imperfection evaluations, and growing, using advanced artificial intelligence (AI) and machine learning to improve its detection rate by nearly 7% at its Smyrna Vehicle Assembly Plant.
The AUTIS Surface Verification System uses high-resolution cameras to capture 15,000 pictures of each vehicle immediately after they’re painted, according to a Nissan news release. The photos are then analyzed to detect possible defects.
Nissan says it uses nearly two gallons of paint on each vehicle during the manufacturing process. That includes three coats (prime, base, and clear), leaving room for imperfections to form on the surface – even in the most sterile environment, the release says.
“The human eye can spot 85-95% of flaws but AUTIS identifies over 98%,” said Travis Fritsche, a Nissan paint process engineer, in the release.
When AUTIS flags a potential blemish, technicians can review the imperfections instantly on monitors and wrist-worn, company-owned smartphones specifically used for the inspection process. The process reduces eye fatigue while increasing the amount of time available to carefully correct the paint, according to the release.
“Technicians are critical for confirming and categorizing defects. AUTIS is a tool to enhance their work, not replace it,” said Rod Lynch, a new model paint engineer at Nissan, in the release. “AUTIS has significantly reduced our investigation time and elevated our quality.”
AUTIS is also used at Nissan plants in Canton, Mississippi, and Aguascalientes, Mexico on Altima, Frontier, LEAF, Pathfinder, Rogue, Kicks, Sentra, Versa, and Murano vehicles.
With machine learning, AUTIS is constantly refining its detection capabilities, the release says.
“It’s a self-learning system,” said Lynch. “With feedback from technicians, AUTIS learns to recognize defects like dirt, fibers, and craters. It will continue to learn and refine and will build a library of what defects look like.”
Nissan says the growing defect library not only helps identify flaws faster but also assists the team in addressing process issues, which are categorized by vehicle type and color.
“AUTIS can group defects to help pinpoint patterns and improve our processes,” Fritsche said.
In 1985, Nissan became the first auto manufacturer to use laser-equipped robots to perform paint inspections, detecting imperfections as small as 0.3 millimeters in 45 seconds, according to the release.
Today, AUTIS can find imperfections as tiny as 0.2 millimeters in less than half the time, representing nearly four decades of technological progress, Nissan said.
“While paint covers the surface of a vehicle, Nissan’s commitment to quality runs much deeper,” Nissan said. “By combining human expertise with cutting-edge AI, Nissan delivers vehicles that meet the highest standards of craftsmanship.”
Images
Secondary photo: The AUTIS Surface Verification System in use at a Nissan plant.
Images provided by Nissan and Business Wire
More information
Nissan creates paint that deflects heat to save vehicle energy consumption