₪... Digital twins

  Digital twins are software or hardware systems that represent a product, process or system in operation or in development. A digital twin can therefore be understood as a simulation/model that is as close to reality as possible. What is new is that when digital twins are in operation, the current state of the system being represented and predictions about its development (under variable boundary conditions) are calculated. These predictions are based on relevant historical data, actual data, expert knowledge and (where relevant) design-related, administrative, legal and procedural specifications/facts.

   Digital twins therefore also represent decision support systems (DSS).

  Particularly interesting variants of digital twins are those that can represent highly heterogeneous state spaces and whose categorization structure can be designed adaptively.

   Adaptive in this context means that:

• new rules are integrated into the old set of rules without delay, without the old set of rules being destroyed or masked.
• the system announces situations that have not been taken into account so farthe system acts in a fuzzy manner, i.e. categorizes in a kind of probability space, whereby situation states can also overlapthe categorizer is designed so compactly that trajectories of the system states can be calculated.
• the categorizer can also be initially conditioned with a small amount of data (states) and only grows gradually during operation (chunking principle).

   These specifications exclude the use of classic neural networks and the method of deep learning, as these cannot represent the game point conditions 1, 2, 4 and 5 listed above.

  IngB RT&S therefore uses the in-house developed CwA principle for its CI-based digital twins, which is based on the actions of the categorizer "cerebral cortex".

   On the next subpages we present in-house developed digital twins from the areas of medicine, technology and maritime condition assessment in the context of the problem of "old ammunition sunk in the sea".

  Of course we can only show a few aspects on this website. Digital twins, especially when they contain heterogeneous data, are complex, very complex. Therefore they contain a structure that corresponds to the user principle "one look - one decision". This of course requires a sophisticated intelligence behind these systems, which we were able to realize through the CwA principle and the cognitive psychology principle of chunking.