From Research to Application: Projects at Fraunhofer EMFT

Gear oils make a significant contribution to minimising friction and thus ensure the safe operation of production facilities. In the Smart Gear project, researchers at Fraunhofer EMFT are developing solutions to detect a drop in the performance of gear oils with the help of sensor technology and machine learning methods and to predict when the oil needs to be changed.

For Bavaria, the agricultural sector has a significant economic importance with a turnover of around 121 billion euros per year. However, conflicts with the economic goals of farmers and the sustainable goals of environmental and animal protection initiatives occur from time to time due to the different interests involved. Especially the emission of ammonia gases leads to social discussions. Researchers at Fraunhofer EMFT are developing a more efficient measurement method for detecting ammonia gases in agriculture. The aim is to contribute to more environmentally friendly and animal-friendly agriculture.

Nowadays, the evaluation of sensor data usually happens in the cloud. However, as a result of ongoing digitalization, the volume of sensor data being collected and analyzed is growing rapidly. In order to transfer the huge volumes of data quickly and securely, researchers at Fraunhofer EMFT are focusing on equipping sensors and actuators with artificial intelligence (AI).

Every year, over 3 million immobile people worldwide suffer from severe, mostly preventable pressure ulcers, with around 0.6 million in Germany. Pressure ulcers not only cause personal suffering but are also time-consuming and costly to treat. Current prophylaxis methods require significant commitment from nursing staff and do not provide continuous security. Therefore, there is an urgent need for individual and continuous monitoring for effective pressure ulcer prophylaxis. SoreAlert, a potential spin-off from Fraunhofer EMFT, addresses this need with a smart patch that allows for automated monitoring of at-risk body areas and provides early warnings for pressure ulcers.

The aim of predictive maintenance is to maintain operating resources proactively and with foresight. This should reduce downtime and maintenance costs to a minimum. Researchers at Fraunhofer EMFT are testing new concepts that use machine learning methods to efficiently process even extremely heterogeneous data and make accurate maintenance predictions for production facilities.

Whether in automobiles – especially in the context of autonomous driving – or future industrial manufacturing: plugs and electrical connection technologies have a key role to play in digital networking. They are the main interface between machines, control units and data processing systems and so they provide the basis for the functionality, simple handling and reliability of automation technology.

In the project ‘Insect cells as sensors for environmental toxins’, Fraunhofer EMFT is developing a novel test method in which living insect cells are used as sensitive biosensors. This allows even the smallest toxic effects of pesticides to be detected quickly, without labelling and automatically. The aim is to identify risks at an early stage and support the development of bee-friendly pesticides – an important contribution to the protection of insects and biodiversity.

The research project »Velektronik - Vertrauenswürdige Elektronik (ZEUS)« of the German Federal Ministry for Education and Research (BMBF), aims to establish a networking platform for trusted electronics as an interface between research and industry.

SENNA is a programmable neuromorphic processor based on a mixed-signal, circuit‑switched Field‑Programmable Spiking Neuron Array (FPSNA).

Edge computing is considered a key for new IoT applications. In order to bring artificial intelligence into future edge products, Fraunhofer EMFT researchers are working together with Fraunhofer IIS and Fraunhofer IPMS as part of the EU ANDANTE project to develop innovative mixed-signal artificial neural network (ANN) accelerator with computation-in-memory (CIM) ability. These are intended to enable the construction of solid hardware and software platforms for AI application development.