Connecting The Industrial Internet Of Things With Smart AI Objects
John Clemons is a Consultant for Rockwell Automation and Maverick Technologies, a leading platform-independent automation solutions provider
Smart manufacturing, Industry 4.0 and the industrial internet of things (IIoT) are discussed constantly in the manufacturing world. As you might expect, some of it’s hype, but a lot of it is more real than you might think. Virtual reality, augmented reality, artificial intelligence (AI), digital twins, additive manufacturing and collaborative robots are just a few of the new technologies making a significant impact in manufacturing.
In the AI world, for example, we already have AI systems processing huge amounts of data in real time, describing what’s going on, predicting what’s going to happen next and even prescribing the actions that need to be taken.
At the machine level, AI is taking another step forward with the marriage of smart objects and the IIoT. That’s happening because even with the IIoT, a good bit of engineering work still has to be done when you’re integrating a bunch of different machines, especially older ones, all together in an IIoT solution.
Smart Objects And Machine Integration
Machine integration can be cumbersome. Engineers must manually search out the data sources, which could mean thousands and thousands of data elements to look through. However, the data may not even be all that easy to find, especially in older machines. Sometimes it’s not even there, and lack of good documentation always makes the task harder.
Then, even when the key data is identified, data integrity and data validity are still big issues and traditional communications approaches all too often cause “noise” in the data. Trying to manage all this makes the task that much more complex.
This is where AI and smart objects come in. Smart objects can act as AI components to encapsulate key aspects of the machine operations such as recipes, configurations, tooling, status and the like, building up these smart AI objects into an object-based AI data model.
This object-based AI data model can become common across all the machines, using it in the IIoT and AI solutions and in the higher-level computer systems and databases. This allows you to use these smart AI objects everywhere the data is needed throughout the entire IIoT.
These smart AI objects not only become a smart way to organize data in the machines themselves but also are standardized across the IIoT making them easy to implement. Tying the smart AI objects to the IIoT allows other machines and systems to easily discover them in the IIoT and allows those higher-level systems and databases to find and use them as well.
How IIoT Leverages Smart Objects
Without the IIoT, the smart AI objects never leave the machine level, and nobody but the machine operators and the engineers even knows they exist.
But with the IIoT, the smart AI objects can be seen and used anywhere and everywhere they’re needed. The IIoT uses the smart AI objects to collect, and then preserve, the context of the data from the machines. Since the IIoT and other systems discover and use the smart AI objects, their context can be created or changed anywhere in the IIoT. This means that edge computing solutions now become so much more practical and so much easier to integrate with the cloud. After all, the data is already encapsulated in the smart AI objects and available through the IIoT.
The IIoT discovers the smart AI objects and their associated datasets, incorporates their underlying organization model, consumes the underlying data, and provides the smart AI objects, the data model and the data, to anything or anyone that needs them — all in real time.
This means you can also embed analytics and machine learning modules pretty much anywhere you want to, using smart AI objects, the IIoT and edge devices. All without being a data scientist.
A Simple Example
As an example, let’s assume we have a few machines that are used to make the parts needed for a finished product.
The smart AI objects exist at the machine level and define the key aspects of the machines, such as tooling, operations, production, waste and alarms. The smart AI objects not only define that data, they also collect that data as well, any time the machine is running. Also, because they’re smart AI objects, they’re well-defined and easily discoverable, so the IIoT can incorporate their data models into the IIoT models. The IIoT can collect the data from the smart objects, making the data available anywhere and everywhere in the IIoT.
Visualization applications like dashboards, reports and analytics get the data from the IIoT and show it to people in real time. Digital twins, edge computing and cloud computing all get the data in real time and incorporate it into their models and calculations.
Using smart AI objects in an IIoT framework isn’t without some challenges. Smart AI objects are in their infancy and very few machine builders are using them yet, let alone even heard of them. Couple that with the fact there are no international standards, or even ad hoc standards, and that means smart AI objects are left up to individual engineers to implement as they see fit.
This means it takes some extra engineering work to define the data models, to define the machine processes to be encapsulated in the data models, and to define how the data models will be used throughout the IIoT. It’s vital to ensure this standard is being followed for all the machines, data models and smart AI objects, throughout the IIoT.
This doesn’t necessarily sound all that glamorous until you realize that these concepts are the very backbone of smart manufacturing and Industry 4.0. Smart AI objects, and the IIoT support everything smart manufacturing and Industry 4.0 are all about.
Smart AI objects and the IIoT also support digital twins and form the backbone of the digital thread. Smart AI objects and the IIoT support the complete convergence of the old tiers of IT and OT, supporting new technologies, platforms, services, applications, roles and capabilities that haven’t even been thought of yet.
That’s another good step forward for AI — connecting the industrial internet of things with smart AI objects.
Discover Past Posts