Autonomy promises new frontiers of efficiency

By Jim Montague

Mar 07, 2019

It was inevitable that automation would lead to autonomy, but it's still unnerving when it happens—like mom and dad worrying when the kids leave the nest. In this case, two videos of a self-docking ship and a self-directed, deep-sea diving vessels led off the "Journey toward Autonomous Industrial Operations" panel discussion at this week’s ABB Customer World in Houston.

Plus, just like those parents, automation and IT providers like ABB and Hewlett Packard Enterprise want to make sure their end users have the essential data they need to run their processes successfully, regardless of how much they're automated or autonomous.

"In a more autonomous world, we need to understand why [these systems] are going onto the factory floor, and how we can provide the continuous analytics and greater collaboration they need," said Häkon Berg, technology development manager, ABB, who added that achieving more autonomy means developing:

  • Flexible technology stacks using containerized software like Docker, integrating speech recognition, and including SMEs in software development.
  • Continuous delivery of code for faster interactions, which can eliminate long product development cycles, perform DevOps that ease application deployment, and enable software to "feel the pain" of plant-floor problems and address its problems more effectively.
  • Platform integration by accessing and integrated data.

Data fuels autonomy

"Whether we're talking about autonomous vehicles or other kinds of autonomy, they all need huge amounts of data and some kind of open cloud for support," added Matthias Roese, chief technologist, Hewlett Packard Enterprise. "This can be complex, such as the car manufacturer we worked with that had to integrate data from 160 vendors on its assembly line, and began using some autonomous functions to avoid doing maintenance earlier or later than needed.

"This can be hard to achieve because, even as some operations become more autonomous, many users still expect to see someone in a blue collar looking at what they're doing. We're also working on a self-healing grid that can show when substations are going up or down, and coordinate power production as they come back up. For example, where it used to take local utilities in Houston three or four days to recover after hurricanes a few years ago, some of the available software apps will soon allow them to recover in just four hours."

Fight fear, follow consumers

Despite the potential benefits, many people remain afraid of artificial intelligence (AI), and the panel agreed these attitudes must be confronted. "There's a lot of fear that AI is going to run the world," said David Funderberg, technology manager, chemicals and refining, ABB. "However, when you go and listen to users, and understand their use cases, you can see how AI can greatly assist them. Give them guidance via devices like HoloLens, help get them to a safe area if needed, and estimate how long tasks will take and what materials are needed to complete them. This isn't closing the loop on control and allowing totally autonomous operations."

Berg added, "One way we're seeking to develop industrial autonomy is by catching up to consumer markets. We're trying to take tools like Alexa and others on the consumer side, and see how they can be reused on the manufacturing side. For example, we're investigating if SMEs can use voice-directed digital assistance to talk to production systems. We're starting small and finding ways to make small steps, but others will come and this will grow and grow."

Hewlett Packard’s Roese concluded, "Because these are big challenges, they'll need a team play to meet them happen. No one company can do it on their own, which is why we appreciate our partnership with ABB."

Subscribe to receive Control eNewsletters

Show Comments
Hide Comments

Join the discussion

We welcome your thoughtful comments.
All comments will display your user name.

Want to participate in the discussion?

Register for free

Log in for complete access.

Comments

No one has commented on this page yet.

RSS feed for comments on this page | RSS feed for all comments