The appeal of a digital twin is obvious: You deploy a digital model that mimics real-world behavior — detecting inefficiencies, optimizing performance and learning to improve in real time.
Digital twins are starting to generate buzz in the tech sector because they represent the next phase in the evolution of computer modeling. Conventional digital models attempt to predict future results before building complex systems, anticipating problems in the virtual realm before they cause expensive trouble in real life. Digital twins build on these strengths by providing instant feedback and machine learning.
Fortunately, the sensors, networks, software and computers required to build digital twins are readily available. Indeed, they’re already finding a home in Industry 4.0 applications, where they’re streamlining production and enabling predictive maintenance.
Moreover, there are strong incentives to move these interactive replicas into new realms. After all, if digital twins can optimize machines, why can’t they improve human outcomes?
That’s where things get complicated.
Consider a sophisticated digital twin used by a professional sports team. It’s conceivable that wireless sensors in clothing and equipment combined with digital video cameras, 3D modeling and advanced learning algorithms could mimic an entire game while it’s being played. Sensors and software could detect which players are off their games and recommend improvements in real time.
This futuristic scenario is easy to envision. Embracing the full potential of digital twinning, however, is a lot more challenging.
Where to Start With Digital Twins
Pretty much any behavior in the natural world is a candidate for digital twin development. It’s not just people and machines. It could be weather, the flow of water or the accumulation of pollutants.
Thus, you’re looking at a veritable mountain of distractions and dead ends. To mine the nugget of productive opportunity with digital twins, you have to narrow your focus.
To do that, ask some basic questions: Do you have a process where you’re reasonably certain that a better understanding of customer behavior will improve revenues or profitability? Do you sell a complex product with a steep learning curve that would benefit from insights on user behavior?
Zero in on a few processes you’d like to optimize and build from there. Remember, machine learning is a core component of a productive digital twin. If you build it right, it will teach itself to work better over time.
It’s imperative to avoid bright-shiny-object syndrome. You have to start with a pressing need — not a nice-to-have.
4 Mandatory Skills for Digital Twin Development
After you’ve targeted your best digital-twin opportunity, you have to amass the skills required to make it happen. These skills span four disciplines:
- Psychology. Like a designer of user interfaces or customer experiences, your digital twin must reflect and respond to cognitive processes that drive human behavior.
- Physiology. A digital twin must account for the subtle interplay of nerves, muscle and bone during the use of a device or software application.
- Architecture. System design drives the effectiveness of a digital twin. Your model must optimize the interactions of networks, sensors and learning algorithms.
- Engineering. Computation combines statistical analysis with the logic of artificial intelligence and machine learning. Savvy engineering supplies the accuracy and relevance that produce success with digital twins.
Your digital twin solution will need equal measures of design creativity, strategic vision and technical acumen. Dovetailing all these skills is a non-trivial task.
Embracing the Future of Digital Twins
Admittedly, digital twin technology is in its infancy. In the years to come, innovations in edge computing and the growth of 5G wireless networks should expand its horizons. Advances in artificial intelligence, machine learning and neural networks will open even more opportunities.
Even if you don’t think digital twins are a good fit for your business today, it’s a good idea to monitor developments — especially in the consumer sector. If a breakout innovation hits the mass market, it could be as big as the smartphone.
We’re not predicting the future at DMI, but we do know promising technology when we see it. As we gather more insight and experience designing and implementing digital twins, we’ll make sure our clients benefit from everything we learn.
–Michael Deittrick, senior vice president digital strategy, chief digital officer