Annual Report: Highlights 2024

Demands on microchip performance, such as for AI applications, are constantly increasing, along with the pressure for those microchips to remain cost-effective. With conventional chip design currently reaching its limits, chiplets promise a viable solution for the future. Thanks to their innovative design, multiple small circuits – each manufactured using the most suitable manufacturing technology – can be combined like building blocks. This optimizes the overall system and makes it possible to create new, more efficient systems in less time, ultimately saving development time and costs. We are working closely with industry partners to advance chiplet technology in Europe. We optimize these modular systems with innovative architectures while also researching the interfaces between chiplets and the power supply.

Together with Fraunhofer IZM and Fraunhofer ENAS, we have founded the pioneering Chiplet Center of Excellence initiative, set up as a research partnership with industry stakeholders. The Center brings together experts from design, tooling, technology, and application across the automotive value chain to ensure the rapid industrialization of chiplet technology. All participants in the initiative share the goal of assessing the feasibility of chiplet-based system solutions at an early stage. In addition, Engineering of Adaptive Systems is working together with the Smart Sensing and Electronics division on a reference platform for chiplets. It’s also leading a work package for the pan-European project Advanced Packaging and Heterogeneous Integration for Electronic Components and Systems (APECS) where we contribute our design and testing expertise to chiplet development.

Under the blazing sun of the African savannah, a vulture glides overhead as it scans the vast expanse below – noting every movement and every shadow. Its goal is clear: when an animal dies, it wants to be first on the scene. This behavior is beneficial for the health of the ecosystem, as quickly removed carrion prevents the spread of disease. Thanks to the GAIA Initiative, these feathered guardians now receive support from AI and satellites. In the GAIA-Sat-IoT project, we’ve developed a transmitter for vultures that includes not only a camera and various sensors, but also an AI that analyzes the collected data directly. Our integrated satellite communication module ensures that the results are transmitted at lightning speed.

Early warning system enables rapid response

GAIA-Sat-IoT combines sophisticated sensor technology with AI and satellite communication to create an ecological early warning system. Wildlife researchers urgently need information on clusters of dead animals, dried-up water sources, and other anomalies to take early countermeasures and prevent ecological crises. The reliability of the information increases when data processing is distributed across multiple transmitters, which connect to each other as the vultures wearing them gather. We explored this concept in another initiative project, SyNaKI. Our findings show that, within the digital swarm, it’s possible to achieve more complex analyses and more precise predictions.

A small chip with a big impact: with our nanoSPECTRAL chip in miniature format, we’ve developed a new technology for spectral analysis. It enables a wide range of applications on objects where conventional spectrometers come up against their limits. Until now, spectrometer technologies have been high-performing, but expensive and unwieldy – making them unsuitable for small spaces and mobile use. With the nanoSPECTRAL chip, which was developed by our Smart Sensing and Electronics division in funded projects, farmers can, for example, determine the ripeness of fruit, doctors can analyze test strips, and cosmetics companies can measure skin moisture.

The nanoSPECTRAL chip shrinks an entire spectrometer down to just 2.3 mm x 3.4 mm – a feat achieved by manufacturing the spectral filters directly in the CMOS semiconductor process, allowing us to integrate them into a chip. Another notable feature of the chip is that its production costs remain almost consistently low regardless of the number of spectral channels, while it quickly and precisely measures the optical spectrum in both transmission and reflection. Thanks to its high level of market maturity, we offer interested companies samples and evaluation kits that they can use to realize pilot production runs of several thousand units. There are no longer any barriers to the future of mini-spectroscopy.

At Fraunhofer IIS’s Development Center X-ray Technology, we’ve been generating large-format tomograms with our XXL CT scanners since 2013. The three-dimensional capture of hidden structures – be they cars, airplanes, or T-Rex skulls – opens up entirely new possibilities. Previously, scanning large, elongated objects required a great deal of effort; to achieve optimum image quality, they had to be positioned vertically using a special device. Our gantry design changes this: in collaboration with the University of Kaiserslautern-Landau (RPTU), we developed an XXL CT system that’s similar to the design of medical CTs.

Instead of being suspended vertically, the object to be scanned remains in its natural position, while the radiation source and X-ray detectors rotate around it. This enables applications that can’t be adapted to the requirements of conventional CT due to factors such as the influence of gravity or the need to apply heavy stresses. Operating at energies of 9 megaelectron volts, this device can penetrate almost any object – and it is the first of its kind capable of capturing details smaller than 100 micrometers in 3D. Fraunhofer IIS is currently building its own gantry high-energy CT scanner, which will significantly expand the possibilities of XXL scanning: in the future, we’ll be able to examine complete vehicles, high-voltage storage systems, and engine components – in pursuit of the most advanced products of tomorrow.

We’ve been setting standards in the automotive audio sector for a decade. In 2014, our Symphoria technology was deployed in Audi production cars, creating a realistic three-dimensional concert hall atmosphere. Since then, numerous manufacturers from the Volkswagen Group have continued this success story. The tried-and-tested Symphoria sound can be heard in many premium models, and our Sonamic technologies are making their way into an increasing number of Audi, Seat, Skoda, and VW models. This year, our audio technologies are also making their debut in BMW and Ford models.

In addition to Symphoria, the high-quality rendering solution for realistic 3D and surround sound in cars, our team developed the Sonamic product range. These technologies ensure an outstanding entertainment experience across all vehicle classes. One solution is Sonamic Panorama, which immerses the listener in the music and creates the feeling of being right at the center of a performance. The technology is already well established on the market and was integrated in some cars with headrest speaker playback in 2024. From this year on, some Audi models have an on-demand-feature called “virtual environments” that drivers can activate if desired. This is made possible by our Sonamic Virtual Environment algorithm, which transforms the car into, say, a jazz club or concert hall.

Our experts are continually refining Symphoria and Sonamic for the vehicles of today and tomorrow. Further research in the automotive sector focuses on the application of edge AI technologies in vehicles. All our solutions contribute to transforming the cars of the future from solely being means of transportation into flexible spaces to work, relax, and be entertained.

In the right place, artificial intelligence can create significant added value for many industrial processes. With Edge AI, it can come to the fore as close as possible to the process. This implies data sovereignty, energy efficiency, and low latency. We’re working on concepts and solutions that use AI directly on existing hardware. A connection to a cloud system is not necessary; it can also work without an internet or mobile connection. The analysis data remains sovereign within the company or the corresponding embedded system.

To make the benefits of AI and machine learning methods available to many small and medium-sized companies, our approach takes two main factors into account: we help our project partners select suitable models, or we develop or adapt the necessary algorithms, and we leverage existing hardware infrastructure. We also employ AI tools that enable super-fast training and adaptation to the companies’ data with minimal manual effort. Ideal applications include machine vision and sensor monitoring tasks, where the focus is on fast response times and high resource efficiency.

Our experts have aligned and optimized our edge AI technologies to make starting digital transformation processes as easy and cost-effective as possible, while also considering energy efficiency and sustainability.

Circular economy is the key to a sustainable industry of tomorrow, as it decouples economic growth from the extraction of primary raw materials. Implementing a circular economy efficiently and successfully calls for digital technologies and data – in other words, a “smart” circular economy, which we are driving forward in our Supply Chain Services division. In addition to new, sustainable business models, companies need a suitable data infrastructure for exchanging product and process data across the entire life cycle. With the digital product passport (DPP), the exchange of this data is gradually becoming a regulatory requirement: it’s used to track the condition and composition of products and to provide information on repair and recycling options.

In the Fluid 4.0 project, we’re working with over 20 partners from the fluid technology industry to implement new business models and data-driven services for a circular economy. Once a shared data space such as Manufacturing-X is in place, we’ll have the infrastructure to share product and process data across the entire life cycle of pneumatic and hydraulic products. The DPP serves as an instrument that will provide static and dynamic product and process information needed to realize the new services and business models that companies are striving for.