Digital representations of the real world are increasingly important for planning, situation assessment and decision-making.
The project PILLARS is the strategic research project for Research Area 3 – “Smart Worlds”, consisting of the groups Semantic Modelling and Acquisition, Geospatial Visualization and Integrated Simulations - all dealing with either the creation or with the handling of digital representations of the real world.
Digital representations of the real world are increasingly important for planning, situation assessment and decision-making. They grow exponentially in volume and complexity and often reside in the cloud. Geospatial and geometric data are complemented by additional data channels such as live and recorded sensor data, simulations and abstract data relevant for planning. This further enhances the overall complexity, resulting in digital surrogates of real objects, which we denote as “smart objects”.
These smart objects further reflect intrinsic properties and interrelationships of their real-world pendants. It is important to make such representations accessible, manageable and comprehensible by applying advanced visual analytics solutions. In our vision, “smart worlds” comprising such smart objects will entirely reside in the cloud, allowing for ubiquitous access and an unprecedented level of distributed collaboration. Moreover, smart worlds will include features for sharing and personalized access control as pioneered by today’s social networks.
To provide smart world solutions fulfilling these requirements, we need to address the following area-specific visual computing research topics:
- Creating smart digital surrogates of real-world objects.
- Semi-automatic decision support.
- Visualization and steering of environmental simulations.
- Uncertainty-aware visual analysis of combined geometric and non-geometric data.
Our strategic research efforts will consequently focus on the following aspects to achieve the required level of dynamism, interactivity and intelligence in these smart worlds:
- Adaptive, self-optimizing visualization and data management systems to cope with real-time data streams, highly dynamic simulations and user interactions.
- Real-time cloud-based processing and visualization strategies to handle large distributed data or to parallelize simulation runs of hundreds of parameterizations.
- Improvement and automation of knowledge extraction during reconstruction, modelling and simulation processes.
- Process mining and automatic workflow improvement based on the analysis of the communication data streams inherent in smart world scenarios.
The implementation activities will base upon the existing, well-developed software platforms, namely Aardvark and Visdom. In all stages of the project, the feedback from applied research projects in the area will be considered.