As a sequel to ’80s cult film Blade Runner is scheduled for US release this fall and will probably reach this side of the Atlantic shortly thereafter, one can’t help but think that companies such as Google have been working with virtual versions of us all along. Those feed on our personal data, and search, browsing and purchase history – no science-fiction here. In fact, in the next few years to come, products and systems will be mirrored in digital twins that will be worth much more than the actual, physical thing. According to Gartner, this is No. 5 of the Top Ten Strategic Technology Trends 2017.
A digital twin is essentially a computerized replica of a physical product, like a steam turbine for example. Unlike a static design model the twin is dynamic – it feeds on continuously updated data and analytics throughout the lifecycle of the physical product.
At last week’s MIT Europe Conference in Vienna, I had the opportunity to interview Dr. Donna H. Rhodes, Director of the MIT Systems Engineering Advancement Initiative (SEAri) and principal research scientist in the Sociotechnical Systems Research Center (SSRC) within IDSS. We talked about Apollo 13 and mirroring systems, ethics, and the fact that, in the near future, you could allow your digital twin to automatically file your tax return.
I read that NASA had started working with this technology as far back as Apollo 13. How far back do you trace the beginnings?
Donna Rhodes: The original meaning of digital twin was strictly a model that as closely represented the product or system as possible. It’s only recently that these models are connected to other devices, and have many sensors, so that you can get real-time feedback.
There’s evidence that NASA in the 70s even was working on this kind of thing. Though their model was not equipped with sensors. Sensors have become a commodity today.
Also the amount of computational power you could have and the amount of data storage you had were limitations. Now that they been removed, we have these enabling environments.
Right now this technology is being implemented in wind mills, steam turbines, aircraft, aerospace – what are growth areas?
Donna Rhodes: Transportation. GE has done work on locomotive and there’s a video online, although I understand that they’re not quite there. Ship-building is another one where there are key advancements. Healthcare is a very active area, as well as an operational digital twin of manufacturing operations or a production floor, and components to products as well. It really varies from a single system like an aircraft to a very complex system like a city or border protection situation, where you have ships, airplanes, and radars.
It depends on what the purpose of the digital twin is and whether you’re relying on it for very deep analytics – like in a health care situation where you’re relying on real time, trying to observe what’s happening in a hospital, and protecting lives. In other cases, you may be looking for patterns in a larger system to anticipate or prevent some kinds of situations.
You said in your talk that the digital model is much more valuable for companies. Do you think that the fear of digital twins being destroyed will lead to having replicas of the replicas?
Donna Rhodes: (laughs) I’m sure that as this moves forward and matures, there will be significant investments in back-ups for these models. Although, if it was something that was considered proprietary, people wouldn’t want to have that in too many different places either…
The impact of destroying the digital twin is going to be much greater than that of destroying the product, because the digital twin will have all of the history and encoded time record of what went on and what people were thinking. If that were destroyed, you probably would never get to that level of completeness.
What is the most important aspect you’d single out in the digital twin technology?
Donna Rhodes: Ultimately the real value of this technology is to anticipate and prevent negative things from happening, and look for opportunities where we can improve something.
The model side is where we can tinker with things without having to stop the system that’s operating. One of the great values here is that we’ll be able to work things out in a world where it will hopefully do no harm, and have a lower risk of making a mistake and not being able to fix it.
What are a few of the ethical considerations?
Donna Rohdes: Anytime you’re studying a situation and observing humans or making value judgements on performance, there might be extenuating circumstances that are not incorporated. The question is: OK, so we have a digital twin of a production floor – are the humans also part of the digital twin and do they also have their digital twins? Where are the limits on this? What does that really mean for the traditional way we reward people for their work or how we try to create a company culture or sense of community?
How plausible is it that my digital twin will be automatically filing my tax return in the future?
Donna Rhodes: (laughs) Some individuals would certainly never want to give up control of this, but there are some others who just don’t want to be bothered. To some extent we’re partly on the road to this. When I do my taxes in the US, every year I use software that knows what I had prior, so there’s already a glimmer of some digital twin of me and my family. Though the term digital twin is maybe not even the right term of what it’s going to evolve in.
The title image is from Vienna’s Albertina Museum Shop that currently has a whole collection of colorful Ottmar Hörl bunnies on display – just in time for Easter. Image credits belong to the author.