Digital Twin means moving an object or environment that has a substance in the real world into a virtual space and linking the two. In simple terms, it means modeling the real world digitally, but it is used not only to model but also to encompass real world data linkage, operation, control, and analysis. There are many similarities to the technology commonly referred to as simulation, but there is a big difference in that it targets a system that already exists in reality, and that it is linked in real time with a real system.
In particular, there is a big difference in that it is linked to reality in real time. For example, if the existing simulation is mainly specialized in what-if, that is, to observe changes when a specific condition is already planned, Digital Twin is Rather, you can observe what's going on in the real world, and you can directly control, operate, and optimize systems that exist in the real world. In other words, if simulation is mainly called design stage, digital twin can be seen as a concept that is used all in design-develop-operation-optimization stage.
The technologies that make up the Digital Twin are very diverse, including modeling, visualization, data linkage, analysis, and simulation. In that it moves the real world into a virtual space, how similar it looks to the real world, that is, the visualization part attracts attention, but in fact, the visualization part is only this part of the elements that make up the Digital Twin. Rather, it can be seen that it is more important to collect various data in the real world, how to link it to the virtual space, how to analyze the linked virtual space, and obtain meaningful insights from it.
The following was presented by LG CNS General Consultant Hee-Kyung Moon at Unity Industry Summit 2019, as it explains in an easy-to-understand way not only the concept of the Digital Twin, but also the case of planning, implementing and operating the Digital Twin for the actual logistics system:
In addition, ETRI is creating and distributing a report on how to use the Digital Twin, which shares related articles and links with reports:
From the point of view of applying AI technology, Digital Twin is also a very interesting field.
- Virtual space is more advantageous than the real world for AI technology application.: Although Digital Twin is a transfer of the real world, it has many advantages in applying AI technology as it is a virtual space that is all represented by digital data. It has the advantage of avoiding many perception errors that may occur when applying AI technology to the real world, and being able to test by randomly assigning much more various conditions than the real world.
- Not limited data set, but advanced real world data: In the process of designing, evaluating, and optimizing AI technology, you usually use learning and evaluation datasets collected under specific conditions, but it is difficult to fully reflect the real world, and it is very difficult to respond to various exceptions. If you use the world built with Digital Twin, real-world data that reflects a lot of real situations will be collected, and you can use this to advance AI technology.
- AI technology is essential for maximizing the usability of Digital Twin: In order to maximize the usability of the built Digital Twin, a feedback loop that quickly and effectively analyzes insights from a huge amount of observations rather than simple observations, optimizes them, and reflects them on the real world system is essential. These analysis and optimization processes are areas where AI technology can be applied very well.
- Provide realistic simulation environment for experimental AI technology: When a product made with AI technology is released, you can pre-distribute it in the built Digital Twin environment, not immediately deploy it to the real world, and perform situation observation and various tests. For example, assuming that you have created an autonomous vehicle, a delivery robot, or a humanoid robot that helps humans, testing it in advance in a Digital Twin environment makes a lot of sense in many ways.
Since the advent of deep learning, AI technology has made a lot of progress, but in reality there are not many cases applied to real-world services. We believe that Digital Twin contains a number of factors necessary to accelerate the application of AI technology, and we believe that synergies in both fields may arise in the future.
