Data Acquisition and Mobilisation

The department's focus areas are derived from these challenges:

• The development of effective methods to ensure data quality in data collection and the reliability and representativeness of information derived from data
• The utilisation of previously untapped data sources by developing new methods for data collection and incentivising data sharing
• The creation of effective and target group-oriented data access and
• The guarantee of secure and trustworthy data supply chains.

The department is dedicated to the consideration of the human factor in the development of data science methods as well as the development of effective concepts for knowledge and technology transfer as elementary foundations for the effectiveness and acceptance of data-based applications.

The department is made up of three groups:

Data Availability and Provisioning

The Data Availability and Provisioning group addresses the department's core challenge, namely the availability of data. To this end, it researches and develops user-centred methods for the acquisition and provision of data, which then serve as the basis for innovative data-based applications and services as well as for data-driven research. Research in data acquisition includes the development of methods for quality assurance in the annotation of data, methods for incentivising data sharing and the trustworthy exchange of data. In the area of data provision, the creation of effective and target group-orientated data access is one of the working group's main focuses.

Secure Software Engineering

Data forms the basis for data-driven business models and research. In this context, it is processed almost exclusively by software, whether in the form of products or data analysis process chains. The development of reliable, robust and secure software therefore plays an important role, especially if this software is a component of complex systems or critical national infrastructure. For example, a software failure or vulnerability to cyberattacks or disruptions can have catastrophic consequences and damage both for satellites and on the ground. It is therefore important to support developers in the realisation of secure, robust and reliable software systems and architectures. The Secure Software Engineering group systematically researches methods, tools and processes for secure software development. Its aim is to automate all security aspects of the software life cycle as far as possible.

It achieves this through contributions to secure design such as automatic architecture-based risk analysis and threat modelling. The implementation is secured, for example, by machine learning-based approaches to programme analysis for vulnerability detection and clone detection. The secure operation of software with complex decision-making is ensured by anomaly-oriented real-time monitoring. Another focus is on securing data supply chains.

Multimodal and Geospatial Information Retrieval

A large proportion of the data available online and locally is not directly usable for the purposes of research, value creation and decision support. The content and location of text documents, websites, images and audio recordings are usually unknown. This information must be identified and extracted using appropriate analysis methods. The group is therefore researching practical methods and developing modular software for collecting, processing, filtering, localising and evaluating the content of unstructured data. The aim is to make previously non-machine-readable data usable for a wide range of DLR applications as a source of information and for decision support. To this end, application-relevant data and information in large amounts of data are robustly identified, made available as required, geolocalised and evaluated with regard to various aspects such as depth of content, topicality, spatial accuracy and reliability.

The application portfolio currently covers the areas of civil security, situational awareness, natural hazards and extreme weather events. In this context, the provision, combination and fusion of location-based text data with geodata and remote sensing data is being researched. In addition, the potential of previously unused text and audio data for applications in the fields of aviation, transport and energy is being investigated.