The enormous volumes of measurement data that obstruct the decision making process in system design is solved by processing and visualising the intrinsic structure of this measurement space without a priori modeling. It is for example known that real inventions emerge, not by hypothising around, but by being surprised of outliers in the dataset.
Traditional static system development has difficulties with modeling real adaptive systems, that have a wealth of sensors and actuators. We solve this by in situ learning of the system in its real context instead of being designed by a human developer.
The customer’s benefits are:
- high level structures are visualized by a 3D map
- detection and visualisation of anomalies in multiple supplementary ways
- combination of exploration (interactive browsing) as well as evaluation services (report generation)low effort since the modeling is in large extent automated,
and in case of complete system modeling:
- The model is visualised by a 3D map that allows model inspection
- Since the model is self learned it allows for automatic adaptation