Core Competence Embedded Software and Artificial Intelligence

The work processes in our society have become very complex and demanding; time is scarce and staff is expensive and hard to get. This is expected to increase further in the future. Electronic assistance systems can support by contributing to a user-friendly, time- and resource-efficient handling of complex tasks. Electronic assistance based on electronic systems with embedded software and artificial intelligence in vehicles is already a matter of course today. In other areas, however, this is not yet so widespread. In the field of nursing care, for example, demographic change in society and the shortage of nursing staff will inevitably lead to the use of automated, technical solutions in order to still be able to cope with the social task of nursing care. Similarly, in the area of industrial production (“Industry 4.0” or “Smart Factory”), electronic systems with embedded software and artificial intelligence can continuously monitor machines and plants by means of sensors, thus ensuring high productivity and quality. In addition to the actual technical functionality an important aspect of such complex system solutions is their operability by the user and the associated acceptance for their use. The use of smart sensor systems at the human interface is indispensable to meet these requirements from the various fields. Smart systems can only exist through the use of embedded software and artificial intelligence.

Communication and Networking

For the connection of sensors to an evaluation unit or the connection of control units to machines, communication and networking are key technologies. Depending on the requirements for data rate, security and latency, different protocols and interfaces are available. Fraunhofer IMS can advise on the selection of these protocols and support with their implementation.

User Interfaces

Almost all modern devices, regardless of their respective purpose and area of application (e.g. industry, traffic, household, hospital), must be configured for their specific task and operated by the user during operation. For this purpose, devices usually have a user interface which is also called "Human-Machine Interface". Fraunhofer IMS offers innovative solutions based on artificial intelligence that ensure intuitive usability.

Machine Learning for Embedded-Systems

Smart sensors offer the possibility of using machine learning methods, e.g. with neural networks, to implement complex data processing and evaluation directly at the point of acquisition and thus to generate high-quality status information from the raw data. This does not necessarily require complex hardware with high computing power; a small energy-efficient microcontroller is sufficient in many applications. This saves energy, space and costs. For this purpose, the Fraunhofer IMS has developed the software framework "AIfES" (Artificial Intelligence for Embedded Systems) which makes artificial intelligence usable on embedded systems. In the course of continuous further development, we provide our partners from science and industry with the latest machine learning methods.

Computer-Vision

The continuous development of new and existing imaging sensor technologies (e.g. CMOS imager, infrared camera or LiDAR) creates the need and possibility for new algorithms in computer vision. Fraunhofer IMS researches try to extract the maximum amount of information from the individual image sensors. In particular, modern methods of machine learning are used. The fields of application range from ultra-fast person recognition over sophisticated motion analysis to the quantification of vital parameters.