Ongoing joint projects:

AIFER: Artificial intelligence for analysis and fusion of earth observation and internet data for decision support in disaster management

Funding codes 13N15525  to 13N15529

Disaster events such as floods, forest fires or storms do not stop at national borders. For this reason, Austria and Germany have decided to join forces and conduct joint research in order to optimally support rescue forces in dealing with major emergencies. The AIFER project is developing a system that uses artificial intelligence to record and evaluate the situation based on the example of a flood scenario. Different data from satellites, airplanes, drones as well as data from social networks will be evaluated, summarized and processed in real time. Legal, sociological and ethical questions will be addressed by the partners.

More information (only available in German)

   

FAKE-ID: Video analysis using artificial intelligence to detect false and manipulated identities

Funding codes 13N15734 to 13N15738

Proving unambiguously the identity of a person is becoming increasingly challenging, as high-quality forgeries of images and videos can now be produced using simple technical means. The goal of the FAKE-ID project is to investigate possible attacks and forgeries of images and videos and to develop a software platform for their detection using artificial intelligence. The result will be a system that can analyze image and video data streams in real time and detect fake or forged material. The detection of deep fakes and the underlying AI decision-making AI remain transparent for the user.

FAKE-ID project outline (PDF, only available in German)

   

KISTRA: Use of AI in early detection of crime

Funding codes 13N15337 to 13N15345

Police crime statistics show a significant increase in politically motivated Internet-related crimes in recent years. Investigating authorities need tools tailored to their needs that enable pre-filtering and support them in criminal prosecution. The goal of KISTRA is to investigate the framework conditions for the ethically and legally justifiable use of artificial intelligence by investigative authorities for the detection, prevention, and prosecution of criminal offenses. To this end, various AI methods for processing data that can be adapted as needed are being developed. The different methods will be brought together and evaluated together with the investigating authorities.

More information (only available in German)

   

KITE: AI-supported VR tactics training for police forces  

Funding codes 13N15546 to 13N15548   

Education and training are essential to optimally prepare emergency forces for real-life situations. Especially in scenarios such as terrorist attacks or rampages, for example at Christmas markets or in shopping centres, on-site training is hardly possible since it would interfere too much with everyday life. Real training environments are costly to create and limited in the number of training opportunities. In KITE, an AI-supported VR simulator is being developed for the training of emergency forces. With it, the training environments can be automatically transferred to VR as 3D models and a wide variety of deployment scenarios can be quickly integrated. 

KITE project outline (PDF, only available in German)

    

KITT: Artificial intelligence to improve the security of tunnels and tunnel control centers

Funding codes 13N15537 to 13N15698

With their bridges and tunnels, highways and federal roads not only form the lifelines for individual and freight traffic in Germany, but they also connect countries with each other. For this reason, Austria and Germany have decided to join forces and conduct joint research not only at the European level but also at the bilateral level in order to further increase the security and availability of roads. In the KITT project, research is being carried out to detect hazardous situations in tunnels at an early stage and adjust the flow of traffic in real time.

KITT project outline (PDF, only available in German)

  

KIWA: Artificial intelligence for flood warning

Funding codes 13N15542 and 13N15543

Dealing with flash flood events after local heavy rainfall is a particular challenge for emergency services. Forecasts of precipitation amounts and runoff processes are hardly possible. Within the KIWA project, AI-based tools for flood warning are being developed and tested. Machine learning-based modelling of precipitation-runoff processes in the field should enable rapid conversion of meteorological forecasts into expected runoff volumes.

KIWA project outline (PDF, only available in German) 

    

KIWaSuS: AI-based warning system for heavy rain and urban flash floods

Funding codes 13N15556 to 13N15560

Due to climate change, local heavy rainfall events are occurring with increasing frequency. They pose a significant security risk to the affected population and a threat to infrastructures and property assets. Torrential rainfall is difficult to forecast, both in terms of location and timing. Especially in urban areas, overloading of sewer networks and rapid flooding often occur. The aim of KIWaSuS is to extend the locally accurate forecast period of local flooding to up to two hours in the event of a storm.

More information (only available in German)

    

CriminalNetworks: Combating billing fraud and corruption in the health sector

Funding codes 13N15530 to 13N15533

The health sector in Germany has an estimated annual turnover of more than 350 billion euros. This arouses criminal covetousness: It is estimated that billing fraud and corruption in the health care system damage the social system by 14 billion euros annually. However, it is not only financial damage that occurs, because patients do not receive the services they actually need due to billing fraud. The CriminalNetworks project is developing an IT tool to support the investigative work of the police.

KriminelleNetzwerke project outline (PDF, only available in German) 

      

UAV-Rescue: UAV-borne sensor technology for AI-based support of rescue missions

Funding codes 13N15549 to 13N15553

Due to events such as gas explosions, natural disasters or terrorist attacks, buildings as well as tunnels can be damaged significantly and people can be buried or trapped. In the UAV-Rescue project, an autonomously deployable unmanned reconnaissance system (UAV) is the main research focus area. The system will be deployed in severely damaged buildings or tunnels to create a complete three-dimensional map of the building interior in real time using artificial intelligence and to detect possible signs of life.

UAV-Rescue project outline (PDF, only available in German)