The concept of digital twin was first conceived by Michael Grieves from the University of Michigan in 2002. However, the first working system was only developed in 2010 by NASA, in an effort to improve their spacecraft via digital representation.
Digital twin, which is a virtual, usually 3D model of a physical entity, has become more relevant due to the technological advances in machine learning that can be applied via digital representation to “real” situations, systems & environments.
The thing is that although these models are undoubtedly useful, they are very difficult to build. So industry struggled to simulate environments, which they needed in order to quickly & effectively find solutions to problems or improve aspects of production or other factors. AWS, understanding this gap in the market, recently introduced the AWS IoT TwinMaker.
Amazon’s vision is to make it simple to build connected digital twins for a variety of use cases that can be updated with live data to drive solutions or conceptualize business outcomes. The key difference between traditional models & digital twins is the “real-time” flow of information, which updates the simulation to accurately offer an exact representation of the physical situation. Let’s look at an example to illustrate the point: consider a packing machine in a large concern that is required to “work” at optimum speed to get the job done, but also to “work” at a conservative pace so that it doesn’t overheat & reduces its own “working life”.
The machine needs to be “calibrated” to produce the best outcome for the manufacturing concern it’s “working” for. Data feed to the machine’s digital twin will gauge its temperature & “feed” it into an AI, ML environment to suggest the best “working” pace to produce the most effective outcome.
Digital twins are challenging to build. An accurate digital model must be first built, it, then, must be “connected” with sensors to the physical entity, data from a variety of sources might be required & visuals must be delivered to the end-user that are useful (usually via the Web).
AWS has developed a variety of tools to ease this process:
- Model Maker – allows users to develop a workspace that offers a visual model of the physical entity it represents. The tool is flexible & “intelligent” enough to create a digital twin graph of a real-world system.
- Components – associate entities with connectors & draw in data from various sources into a central store on the workspace.
- Scene Composer – console-based 3D scene composition tool that supports previously built 3D CAD models (which the system converts to a usable format). Visual assets can be “placed” to represent their “real-world” counterparts, tagged & connected to useful data, such as operating manuals & so forth.
- Applications – provides a plug-in for Grafana & Amazon Managed Grafana to create dashboards. 3D scenes can be embedded on the dashboard. Other useful tools, such as a video player, hierarchy browser, time-series data charts & tables can also be added. The dashboards use AWS IoT TwinMaker’s unified data access APIs to populate the widgets used.
To know how to get started with the AWS IoT Twinmaker, click here.