“Characteristic sounds and vibration patterns are often the only clues for an optimal separation process or necessary readjustment,” says von Enzberg. INTELLIGENT DISTRIBUTION Centrifuges play an important January 2019 Calendar Telugu role in the industrial production of foodstuffs. Until now, the experience, hearing and intuition of a machine have been optimally adjusted Operator required. Now, it has been possible to transfer this knowledge to an expert system using neural networks.
January 2019 Calendar Telugu
FRAUNHOFER IEM25Complex sensors and intelligent data processingThe installation of sensors that make the drain technically more manageable, is complex and demanding, as they have to provide data of the machine continuously in the millisecond range, which additionally be measured at different places during the rotation. The research team at the Fraunhofer IEM has now mounted sensors not only on the housing and drainpipe, but also inside a rotating container – these can make and transmit up to 48,000 measurements per second.
But the abundance of this data must also be processed quickly and intelligently, so that the centrifuge are controlled as real-time as possible can. »Aim of the intelligent data It’s about understanding the perception and judgment of a human operator, “emphasizes von Enzberg. “Therefore, we rely on machine learning and have used neural networks and the number of data obtained to train one of our programmed systems. It can detect fault conditions and develop strategies for adjusting machine parameters to fine-tune the machine even during operation in fractions of a second.
“This knowledge can also be applied to the handling of different centrifuges and tasks. This requires that the machines and systems are equipped with appropriate sensors and that the database is large enough. Once the learning process has been completed, centrifuges will in future also be able to work autonomously and remotely.www.iem.fraunhofer.de26 “MCube” microscopy system According to WHO estimates, around 214 million people worldwide were affected by malaria in 2015 – with approximately 438,000 deaths.
The disease is caused by parasites of the genus Plasmodium, which can be detected by microscopic examination in blood smears. The microscopic examination, however, can be very time consuming if there are only a few exciters in the sample and a large number of fields of view are to be assessed manually. As part of a multidisciplinary project of the Fraunhofer Future Foundation, the Fraunhofer Institute for Integrated Circuits IIS is therefore developing a computer-aided microscopy system for the automatic recording of blood smears and detection of malaria pathogens based on Artificial Intelligence methods.
Www.iis.fraunhofer.deTRENDS & PROJECTSView the Black Box Today’s AI models and machine learning are large (gigabytes), complex – and thus energy hungry both execution. Consequently, the models can be executed on high-performance computers, but not on embedded devices, IoT devices or smartphones. The Fraunhofer Institute for Telecommunications, Heinrich Hertz Institute, HHI develops techniques to reduce the complexity of neural networks and to compress them without sacrificing performance .
In machine learning in critical applications, methods are also developed to be able to study neural networks more deeply – for example, to verify that the solution path of the AI is meaningful and that it January 2019 Calendar Telugu comes to correct results on the right path: the so-called “black box” look .www.hhi. fraunhofer.de27 Optimized energy networks By 2020, thousands of kilometers of new grids are to be created in Germany in order to make use of renewable energies.