Disrupting the Production Line with AI Edge Inference

February 26

In the manufacturing industry, some of the most closely guarded secrets revolve around production line innovations. In the race to ultra-efficiency, companies are experimenting with new Industrial Internet of Things (IIoT) solutions that will help keep them miles ahead of the competition.

Artificial Intelligence (AI) chipsets for edge inference are enabling a large portion of these new technologies. The processing of unstructured data closer to the data source allows a reduction in latency, optimized performance, and faster response times as well as significantly reduced costs for data transfer.

As mentioned in a previous blog post on edge computing, only companies with the deepest pockets can participate as pioneers of technological development. German automakers are using robot analytics by Software AG and Dürr to imagine a futuristic paint shop that can accommodate fluctuations in market demand. In the new factory layout, traditional lines have been broken up into shorter sequences to easily adapt production capacity for a variety of model ranges and painting processes. The ability to identify and correct defects in real-time without affecting the entire production line saves time and material while increasing the overall quality of the final product.

This type of Industry 4.0 overhaul is not a one and done procedure, and rollouts can only hit one factory at a time. Implementation and training must go hand in hand for it to be successful. Analysts forecast revenue from the sale of AI chipsets for edge inference and inference training to increase dramatically between 2018 and 2023. While shipment revenues from edge AI processing units reached 1.3 billion in 2017, they are expected to climb to $23 billion by 2023.

Killer Apps for Quality Assurance

In 2019, Bühler Group won the innovation award at VICTAM International for its optimized inspection system, which can detect and remove contaminated grains from the production line. Aflatoxin is a big challenge for crop growers because direct or indirect consumption of grains affected by the mold is known to cause liver cancer and growth deficiencies in children. Incredibly persistent, this dangerous micro toxin can even be transmitted through milk from cows exposed to the contaminated grain.

This video shows how Bühler’s technology uses light and air to identify and expel defective grains, seeds, nuts, or plastics from the production line. Their award-winning LumoVision sorter uses UV light and an AI-enabled camera to identify microscopic defects at rates of approximately 10-15 tons (about a truckload) in as little as one hour. While human eyes are not suited for finding errors in large scale, fast-moving production lines, edge systems combined with AI technology are bringing hyper-efficiency to the quality assurance process.

Leading Enterprise to the Promise Land

Another group leading Industry 4.0 technology out of Germany is the ADAMOS (ADAptive Manufacturing Open Solutions) collective. Founded by Dürr, DMG MORI, and Zeiss, they have created an open-source IIoT platform and digital marketplace for mechanical engineering applications. Developers and machine tool builders can use the manufacturer-neutral IIoT platform to bring scalable, IoT dashboard capabilities to companies with legacy machines and equipment.

Ralf Dieter, CEO of Dürr Group and Chairman of ADAMOS, was recently interviewed and asked about the challenges he sees facing Industrial IoT adoption. Despite their successes in the space, Dieter found that many companies were slow to sign with them. When pressed to disclose their hesitation, these decision-makers admitted they did not have a digital strategy in place to facilitate usage of the technology. While the company had created a robust framework for developer support, its enterprise customers were still facing a considerable knowledge gap.

To address this need, ADAMOS created a unique digital transformation program to provide training, consulting, and implementation services. By providing a framework for implementation in addition to their product, ADAMOS was able to bring peace of mind to enterprise clients who are already nervous about digital transformation.

Painting the Bigger Picture

ADAMOS provides an essential lesson for IoT ecosystem partners to remember when approaching clients. Although many C-suite executives are eager (or resigned) to digital transformation, it is a hard learning curve to undertake while simultaneously running the day-to-day operations of an existing company. While not every Industrial IoT solution will spark industry disruption or require a blockbuster budget, to get buy-in from new clients on a solution, it is crucial that they understand the full picture and not just one piece of the puzzle.

Jennifer Davis

About the author

As a highly-skilled technical writer, Jennifer brings practical knowledge and experience communicating solutions around green tech, aerospace, blockchain, LPWA, and Mobile IoT applications in a variety of industries and vertical markets. Prior to branching into technology, she gained over a decade of experience developing creative content for print and digital media.


{"email":"Email address invalid","url":"Website address invalid","required":"Required field missing"}