Research project SPELL launched in June

Mastering Crises with Artificial Intelligence

In June 2021, the research project “SPELL” was launched. Funded by the German Federal Ministry of Economic Affairs and Energy, it will run for three years. The aim of this project is to enable emergency response, emergency aid, and supply measures for the population to be initiated more quickly and in a situation-appropriate manner in crisis situations. This is to be achieved with the help of Artificial Intelligence (AI). One of the project partners is Fraunhofer IESE from Kaiserslautern, which is responsible, among other things, for the topics “AI-supported value-added services”, “Evaluation of the platform”, and “Data usage control in the ecosystem”.

© iStock.com/ipopha

“SPELL” stands for “Semantic platform for intelligent decision and operations support in control centers and situation centers”. In crisis situations such as major incidents, pandemics, natural disasters, or widespread power outages, resources must be distributed correctly and the best decision must be made on the basis of the available information. A prerequisite for this is a data-based overview of all relevant information as an overall picture of the situation, and the networking of all those involved. This is also demonstrated by the Covid 19 pandemic currently underway. Artificial Intelligence can create this basis and significantly support decision-making and implementation in order to minimize wrongdecisions.

The contribution of Fraunhofer IESE
The project SPELL is thus about building an ecosystem for AI-supported crisis management. Here, Fraunhofer IESE is responsible for the dependable assurance of the planned AI-supported value-added services and for the evaluation of the platform. AI-supported services process incoming sensor data to predict, for example, the spread of pollutant clouds. Due to the resulting consequences, the Machine Learning methods applied in this context must always deliver reliable results. This is where Fraunhofer IESE comes in with its methodology, providing statements with respect to the dependability of the expected results of the algorithms. 

In the course of the evaluation, which is a standalone work package in the project, the requirements of the project itself as well as those for the evaluation will be elicited first, and information about the (potential) users of SPELL will be compiled. In the further course, the technical implementations will be evaluated from the perspective of the users. In addition, a didactic design will be developed to enable targeted training and instruction measures for the users.

Some of the data fed into the platform is security-relevant, valuable, relevant in terms of data protection law, and ultimately belongs to the data donor. For this reason, the “MY DATA Control Technologies” framework developed by Fraunhofer IESE will be part of the research project. The developed software will be primarily used to assess, enforce, and manage security and privacy regulations. For example, the owner of the data can specify exactly which platform service is allowed to use their data for what purpose and for how long.

The project partners
Twelve partners with different focus areas are collaborating in this project: German Research Center for Artificial Intelligence (consortium leader), Advancis Software & Services, apheris AI GmbH, BASF SE, Corevas GmbH & Co., DRK-Landesverband Rheinland-Pfalz e.V., Empolis Information Management GmbH, Fraunhofer-Gesellschaft e.V. with its Institutes IESE and FOKUS, ISE Informatikgesellschaft für Software-Entwicklung mbH, LiveReader GmbH, Technische Universität Darmstadt, VfS Verband für Sicherheitstechnik, and other associated partner organizations.