SDF | Sensor Data & Information Fusion

Even before any scientific consideration or technical implementation takes place, sensor data and information fusion is an omnipresent phenomenon: every living creature merges impressions of different, complementary sensory organs with previously acquired empirical knowledge and messages from other living creatures. It uses the result to create a model of its environment - the basis for situationally appropriate action.

As a branch of applied computer science, sensor data and information fusion attempts to comprehend this information linkage: automate it as far as possible and enhance it beyond natural capabilities. In this respect, it is a branch of automation technology, a kind of mechanical engineering for "cognitive tools" to enhance the human abilities of data evaluation and fusion.

Sensor data fusion takes advantage of general technology trends such as networking, mobility, sensor and database technology and creates a basis for effective interaction between people and the technical systems that support them. Based on decades of experience, the Sensor Data and Information Fusion department offers research and development, from conceptual studies to the integration of system prototypes.

Distributed sensor systems

Multi-sensor situational pictures are indispensable in all areas of reconnaissance and surveillance, including threat analysis and the protection of people and infrastructure. The Distributed Sensor Systems research group develops methods for the fusion of different sensor data collected in distributed sensor systems, in particular in multi-static sensor systems. It also sets up experimental systems to evaluate the performance of different approaches under real-world conditions and to identify points for optimization.

Sensor and Resource Management

The Sensor and Resource Management research group establishes methodological and algorithmic foundations for adaptive and optimized management of multi-sensor systems, multifunctional sensors and mobile sensor platforms. Characteristic of its work is the interlinking of methods derived from mathematical optimization theory with data fusion at the signal and data level. This enhances reconnaissance performance without the need for new investment in hardware. Conversely, fusion at a higher processing level can increase the power of the signal processing chain.

Array signal and multichannel processing

Innovative sensor concepts and advanced signal processing methods enable new sensor functions and a high degree of measurement accuracy. Precise sensor modeling is beneficial for the fusion of the measurements obtained. These models enable complex system simulations, assessment of system performance and derivation of technical system requirements. In addition to emitter localization, methods of array signal processing also offer solutions for robust navigation in the presence of jamming and deception.

Integrated sensor systems

To solve demanding reconnaissance tasks using a combination of different sensors, hardware concepts are created and added value is tapped through innovative algorithms. Unmanned aircraft systems (UAS) are the target platforms for this research. Another focus is ground-based, multi-sensor short-range reconnaissance, particularly for the detection of unmanned aircraft systems and dirty bombs.

Infrastructure

By maintaining a laboratory for the development and design of demonstration systems and subsystems, the Sensor Data and Information Fusion department is capable of testing and demonstrating prototypical implementations of the latest algorithms for processing and controlling sensor data directly under near-real-world conditions.

Official bodies/networks

• NATO activities
• UAV DACH
• Fraunhofer SPACE Alliance
• Fraunhofer Big Data Alliance

Collaboration

• University of Bonn
• TU Ilmenau
• RWTH Aachen
• University of Applied Sciences Bonn-Rhein-Sieg
• University of Connecticut (USA)
• Koblenz University of Applied Sciences
• Technical University of Munich
• University of Siegen