Spectrum analysis and vector network analysis are enabling technologies for component development throughout the microwave and millimeter wave band. Due to the lack of affordable electronics for frequencies above 100 GHz, vector network analyzers (VNAs) have to use frequency extenders to reach into the THz frequency band (100 GHz-10 THz). The bandwidth of frequency extended electronic systems is restricted to about 50%. Several extender setups have to be used for larger spans, requiring realignment and tedious recalibration. Further, extenders become increasingly expensive the higher the THz frequency and are, therefore, barley used. Electronic spectrum analyzers face similar problems as VNAs. Only a few examples of highly expensive photonic, pulsed frequency comb-based systems have been demonstrated. Affordable, large bandwidth commercial THz metrology tools are missing so far. This proposal aims for the development of photonic THz characterization tools based on telecom-wavelength compatible photomixing technology to satisfy this need: 1.) Photonic vector network analysers (PVNAs) with extreme frequency coverage will be realized by two approaches: a) A planar, on-chip, and broadband dielectric waveguide topology with integrated photomixers for the realization of a continuous-wave (CW) two-port PVNA, covering at least 100 GHz to 1.1 THz in a single setup. b) A pulsed, free space photonic two-port VNA for frequency extension towards 5 THz. 2.) CW photonic THz spectrum analyzers (PSAs) with a frequency coverage of at least 50 GHz -1.1 THz and a simple extension towards 2.7 THz. This system will be realized both free space and on-chip by using a photonic sweep oscillator that is mixed with the signal to be investigated and down-converted by a room-temperature operating THz detector. These systems will provide a solid basis for THz component development. The long term goal beyond this proposal is a competence center for THz device engineering.

Present-day industrial robots are made for the purpose of repeating several tasks thousands of times. What the manufacturing industry needs instead is a robot that can do thousands of tasks, a few times. Programming a robot to solve just one complex motor task has remained a challenging, costly and time-consuming task. In fact, it has become the key bottleneck in industrial robotics. Empowering robots with the ability to autonomously learn such tasks is a promising approach, and, in theory, machine learning has promised fully adaptive control algorithms which learn both by observation and trial-and-error. However, to date, learning techniques have yet to fulfil this promise, as only few methods manage to scale into the high-dimensional domains of manipulator robotics, or even the new upcoming trend of collaborative robots. The goal of the AssemblySkills ERC PoC is to validate an autonomous skill learning system that enables industrial robots to acquire and improve a rich set of motor skills. Using structured, modular control architectures is a promising concept to scale robot learning to more complex real-world tasks. In such a modular control architecture, elemental building blocks – called movement primitives, need to be adapted, sequenced or co-activated simultaneously. Within the ERC PoC AssemblySkills, our goal is to group these modules into an industry-scale complete software package that can enable industrial robots to learn new skills (particularly in assembly tasks). The value proposition of our ERC PoC is a cost-effective, novel machine learning system that can unlock the potential of manufacturing robots by enabling them to learn to select, adapt and sequence parametrized building blocks such as movement primitives. Our approach is unique in the sense that it can acquire more than just a single desired trajectory as done in competing approaches, capable of save policy adaptation, requires only little data and can explain the solution to the robot operator.

Twenty years of research in magnetocaloric materials has failed to provide the necessary breakthrough that will lead to a commercial realisation of this technology and satisfy the urgent global need for more efficient refrigeration. We strongly believe that this is a result of looking in the wrong direction. The cool innov project will achieve this breakthrough by rethinking the whole concept of caloric cooling. We are rejecting the conventional idea of squeezing the best out of magneto-structural phase-change materials in relatively low magnetic fields, and instead we introduce a second stimulus in the form of pressure so that we can exploit, rather than avoid, the hysteresis that is inherent in these materials. The hysteresis will allow us to lock-in the magnetisation at saturation as the magnetising field is removed, so that magnetic fields persisting over a large area will no longer be required (instead, we can use a very focused field), and then demagnetise the material in a second step with an applied stress, enabling us to extract a lot more heat. In this case we only need to apply the magnetic field to a small volume of material, making it a completely new application for commercially available, high-temperature, YBCO-type, bulk superconducting permanent magnets. With the high-field, multi-stimuli approach proven, we will develop new magneto/mechanocaloric materials that match the new high-field, hysteresis-positive approach and start to fabricate novel heat-exchanger structures using additive manufacturing, so that we can combine a mechanically sound heat exchanger having a complex geometry with locally tailored, magneto/mechanocaloric properties. The success of cool innov will be game changing. We are being very ambitious in targeting a revolution in cooling technology, but if we succeed, we will have a huge impact on global energy consumption through greater efficiency, thanks to the novel energy materials that will be discovered within cool innov.

Today, when using services on the Internet, users have to fully entrust a single service provider with their data. Many of these service providers are located outside the EU and there are cases where data has not only been leaked by attacks of outsiders or insiders, but also by governments who built backdoors into software or hardware, or forced service providers to give out sensitive user data. With the new EU General Data Protection Regulation (GDPR) also companies have an obligation to properly protect users’ data. My project PSOTI will eliminate the need to trust a single service provider and empower users to freely control their data. For this, the users can choose a subset of multiple service providers that they are willing to trust who jointly process their data and privacy is guaranteed even if all but one are compromised. The main goal of PSOTI is to develop privacy-preserving services for commonly used tasks on the Internet that are feature-rich and efficient enough for practical use. This will allow to privately store, retrieve, search, and process data, and help to comply with the GDPR and preserve the fundamental rights to privacy and the protection of personal data. As underlying technology, we will build a real-world secure multi-party computation (MPC) framework that can also be used for other large-scale privacy-preserving applications such as genomics or machine learning. To achieve our main goal, we will solve the following challenges: 1) Develop private query protocols on outsourced data that process complex queries such as Boolean formulas over string matches or range queries, and even hide the query’s structure. 2) Build a real-world MPC framework that scales to large functionalities, is highly parallelized, interoperable, and fully integrated. 3) Demonstrate real-world applicability for privacy-preserving and feature-rich services on the Internet such as file storage (going beyond Dropbox), surveys (going beyond Google Forms), and email.

Understanding oxygen dynamics is a key to superior device performance in emergent oxide electronics. So far it is an unrealized dream to correlate electrical behavior and atomic structure during device operation. Here, I envision bridging the gap between theoretical models and experimental reality. Recent advances in microelectromechanical systems (MEMS) chips for in situ transmission electron microscopy (TEM) are opening exciting new avenues in nanoscale research. The capability to perform current-voltage measurements while simultaneously analyzing the corresponding structural, chemical or even electronic structure changes during the operation of an electronic device would be a major breakthrough for nanoelectronics. Controlled electric field studies would enable an unprecedented way to investigate metal-oxide functional devices by using a lab-on-a-chip approach. I propose this project based upon own groundbreaking work on (i) how to electrically contact and operate an electron transparent lamella device fabricated from a metal-insulator-metal (MIM) structure (Ultramicroscopy 181 (2017) 144-149) and (ii) the design of a novel MEMS-based chip for in situ electrical biasing. FOXON will provide a platform for atomic scale operando investigations of oxide thin film and interface switching phenomena of MIM devices under electrical bias inside a microscope. My scientific endeavor will establish a group to develop beyond state-of-the-art operando TEM of MIM structured devices and tackle open questions in the field of oxide electronics. My scientific mission incorporates (a) studies of switching processes in oxide devices and (b) a comprehensive understanding of the atomic-level mechanisms that lead to tunable physical properties including dynamics of oxygen vacancies and stability of quantized conductance states in RRAM devices (Adv. Funct. Mater. (2017) 1700432). The results from this ERC Starting Grant could pave the way for novel quantum and information technologies.