In 1913, Henry Ford revolutionized the automotive industry with the first movable assembly line, an innovation that made assembly of new vehicles faster and more efficient. Nearly a hundred years later, Ford is now using artificial intelligence to speed up existing manufacturing lines.
The Ford plant in Michigan, where robots help assemble torque converters, now includes a system that uses artificial intelligence to learn from past attempts how to move parts in place most efficiently.
Ford is using technology from a start-up called Symbio Robotics for sensing and controlling the arms, which is looking at hundreds of past attempts to determine which techniques and motions seem to work best.
Toyota and Nissan are using the same technology to improve the efficiency of their production lines.
Technology allows this part of the assembly line to run 15 percent faster, which is a significant improvement in automobile manufacturing where slim profit margins are largely dependent on manufacturing efficiencies.
The company plans to explore whether it uses the technology in other factories, as the technology can be used wherever the computer can learn from the sense of how well things fit together, and there are a lot of these applications.
AI is often seen as a transformative technology, but the assembly of torque converters at the Ford plant illustrates how AI can infiltrate industrial processes in incremental and often imperceptible ways.
Although the auto industry has become highly automated, the robots that help assemble, weld and paint vehicles are powerful and accurate machines that repeat the same task endlessly with no ability to understand or interact with their surroundings.
Adding more automation is challenging, and includes jobs that are still out of reach for task machines, such as adding flex wires to through the car’s dashboard and body.
And in 2018, Elon Musk blamed the decision to rely more on automation in manufacturing, causing Tesla to delay production of the Model 3.
Researchers and startups are exploring ways to give robots more capabilities, for example enabling them to perceive and understand unfamiliar objects moving along conveyor belts.
The Ford example illustrates how existing machines can often be improved by offering simple sensing and learning capabilities.
Artificial intelligence is increasingly used for quality control in manufacturing, as computer vision algorithms can be trained to spot defects in products or problems in production lines.
One of the main challenges is that each manufacturing process is unique and requires automation to be used in specific ways, and new technology must be integrated into the workflow without compromising productivity.
