Stephan Pottel, Manufacturing Practice Lead, Zebra Technologies explains how cars parked on streets attached to charging cables is fast becoming a common sight with the shift from combustion to electric never out of the headlines. And it seems some carmakers have already moved beyond the transition phase.
For example, Walter Mertl, Chief Financial Officer for BMW, revealed that the carmaker now generates more sales growth from electric vehicles (EVs) than combustion. But it’s a journey, and the speed of change varies for each carmaker and region.
With about 100 million lines of code and 1,000 or more semiconductor chips, the digitisation and computerisation of cars has been another major shift witnessed over time and it keeps on growing. Today’s autonomous vehicles are estimated to have around 300 million lines of code, and an electric vehicle needs about 3,000 chips.
And a new layer or transition might already be underway as cars, such as those with Volkswagen’s IDA voice assistant that now come with ChatGPT, while BMW has partnered with Amazon to introduce large language models (LLMs) into BMW’s Intelligent Personal Assistant. At this stage, it’s too early to say how generative AI will change the way drivers, cars, and surroundings interact with each other in the long term, but it’s exciting to imagine the possible use cases and experiences beyond those presented by BMW and Volkswagen.
Drivers could benefit from a personalised AI assistant to help with planning trips, setting reminders, finding space in a car park, sharing journey status with others, delivering real-time route updates and recommendations. It could even order coffee or lunch, ready to collect as you arrive at the drive-thru of the next service station.
From plant to electric car – more AI in auto manufacturing
The introduction of generative AI into cars signals the growth of AI as a value driver for consumers in automotive manufacturing, but they’re not the only ones to benefit from AI— manufacturing engineers are also gaining an advantage. There’s a number of AI applications that carmakers are capitalising on to meet quality and compliance requirements for modern cars that people want to drive as well as making the plant worker experience better. One type of advanced AI is deep learning which imitates human neural networks found in the brain.
Academic researchers and technology companies are turning their attention toward deep learning as a useful tool for automotive manufacturers faced with a variety of visual inspection requirements that traditional tools have struggled to handle. Traditional machine vision systems are used for quality and end of line inspection, traceability of parts, gauging and measurement, presence/ absence checking, metrology and porosity inspection. However, these tools come with longstanding problems, including training time needed, cost, interoperability, maintenance, and handling complex use cases.
But things are changing. In a recent Zebra industry benchmark report, 56% of automotive business leaders in the UK and 43% in Germany said they are currently using some form of AI such as deep learning in their machine vision projects. And an average of 20% in both the UK and Germany say they are not using any AI but would like to know more or are currently looking to procure.
Deep learning machine vision “AI eyes” secure previously unobtainable levels of accuracy, quality, and compliance checking—and can bring engineers, programmers, and data scientists together with new deep learning tools for modern car production.
Deep learning machine vision can be used in high-demand semiconductor production, from wafer inspection, pattern alignment, die sorting, wafer dicing, solder paste quality, metrology and 3D inspection. High standards are needed more than ever to power everything from cars to generative AI to cars with generative AI. Deep learning machine vision can benefit the EV battery making process too, enhancing the inspection of node and cathode coating, electrode tab position, stacking alignment, serial/code inspection, and assembly verification, as the industry advances its electrification efforts.
And at a time when manufacturers face challenges with hiring and retaining skilled workers, ready out-of-the-box deep learning tools are a game-changer. A deep learning optical character recognition (OCR) tool can come with ready-to-use neural networks pre-trained using thousands of different image samples, delivering high accuracy straight out of the box, even when dealing with very difficult cases. Users can create robust OCR applications in just a few simple steps—without the need for machine vision expertise. It’s an example of how advanced AI tooling can take a low/no code approach, so AI is democratised and easier to use for more workers.
In the next five years, 33% of decision-makers in the UK automotive industry and 29% in Germany want to automate over half of their visual inspection processes using machine vision. That goal will be a struggle to achieve without modern machine vision.
AI, whether deep learning or generative AI, is a value driver for consumers and those, such as front-line plant engineers, responsible for moving manufacturing forward. AI is creating new levels of asset visibility, better informed and equipped workers, and expanding the realm of the possible when it comes to automating repetitive and complex tasks. Valuable, skilled engineering teams can act as humans-in-the-loop, while giving more time to moving manufacturing strategy and operations forward toward whatever transition comes next.
Learn more about deep learning machine vision here.
For more stories on Industrial Data & AI click here.