Kurt Fuchs, HI ERN

Perovskite solar cells might become a sustainable alternative to conventional silicon-based solar cells. In the journal Science, researchers around Pascal Friederich, KIT, and Christoph Brabec from the Helmholtz Institute Erlangen-Nürnberg (HI ERN) now demonstrate a closed-loop workflow that combines high-throughput synthesis of organic semiconductors with device characterization and Bayesian optimization to discover new hole-transporting materials with tailored properties for solar cell applications. The predictive models were based on molecular descriptors that allowed to link the structure of these materials directly to their performance in solar cell devices. A series of high-performance molecules, identified from minimal suggestions, achieved up to 26.2% (certified 25.9%) power conversion efficiency.

KIT

Waves are all around us. To understand them is to understand nature. For about ten years, the Collaborative Research Center (CRC) “Wave Phenomena” has analyzed water, sound, pressure and electromagnetic waves, as well as abstract phenomena related to wave propagation. Waves are too diverse and complex to be defined generally, but they share a common property: a temporal change is always associated with a spatial one. "Partial differential equations help us to describe this phenomenon by coupling changes in time and space," says Björn de Rijk. The tasks of the CRC include developing new numerical methods as well as testing and improving existing methods. The goal is to eventually control waves under certain conditions for technical or medical applications – from mobile communication to pacemakers, explains Benjamin Dörich.

Amadeus Bramsiepe, KIT

Robotic gripping systems are an important component of automation technologies - in industrial production, logistics or medicine. Researchers from KIT and the University of Stuttgart are jointly investigating how a robotic gripping system can learn by imitating humans and to react quickly and flexibly to changing requirements in the new HaptXDeep future lab. "We are using autonomous imitation learning and deep reinforcement learning for our system," says Rania Rayyes, project manager at KIT. At the University of Stuttgart, colleagues develop software for the safety and reliability of the robot, which will then be tested at KIT. In the future, HaptXDeep will also include projects on sensor technology, gesture recognition via motion tracking and the control of different fingers in robot gripper arms.