CWTe: Manufactured Electronics for Microwave Applications & New Design Thinking

Please be invited for the February 6 edition of the Center for Wireless Technology (CWTe) bi-monthly^* colloquium webinar,

Webinar call-in details:

________________________________________________________________________________

Click here to join the meeting

We will have two speakers in this colloquium/webinar on interesting wireless topics:

Rafael Del Rey, Nano Dimension – Additively Manufactured Electronics for Microwave Applications & New Design Thinking

Bram van Berlo, TU/e M&CS – Towards Domain Invariant Block Avoidance: mmWave Recognition Feasibility and OFDM/Radar Dataset

Speaker 1:

Title Additively Manufactured Electronics for Microwave Applications & New Design Thinking

by Rafael Del Rey (Nano Dimension)

Abstract

The need for more complex but sustainable electronics makes electronic development an increasingly convoluted task. New manufacturing tools and methods are thereby required to generate new sustainable proposals without resigning on complexity, such as Additive Manufacturing of Electronics (AME).

Fully three-dimensionally printed electronics bring entirely new possibilities to the design and application of such products by enabling complex geometries, embedding 3D-printed components, and significantly shortening the development cycle. Furthermore, AME enables embedding of passive and active components within a single package/board.

Nano Dimension uses a manufacturing process (inkjet printing) in which conductive silver nanoparticle ink and insulating ink are bonded and cured/sintered layer by layer, allowing new geometric freedom in 3D and thus a variety of novel applications (RF, IC staking, encapsulated sensors, unconventional system-in-package solutions) allowing them to include complex features such as curved vias, coaxial and waveguide transmission lines or truly twisted pair routings.

This presentation will showcase the latest developments of AME technology within Nano Dimension and how Nano Dimension foresees the future design of printed circuit boards, starting with the status quo on current PCB technologies, the generalities of the AME process, and the AME potential applying a new design thinking for RF applications, multilayer and complex high density interconnects (HDI, non-planar transmission lines, embedding & encapsulation of components, and electromechanical printed devices, finishing with ways to achieve it using current 2D ECAD tools along with 3D MCAD.

With nearly two decades of expertise in High-Speed Hardware Design, Dr. Rafael del-Rey is currently Global Application Engineering Director at Nano Dimension.

Formerly, as lead system engineer at Volocopter, he spearheaded the electrical architecture of an experimental aircraft. He has also contributed significantly to high-performance computing hardware research at Continental Automotive; developed silicon reference designs at NXP, and also mission-critical server platforms at Intel.

Dr. del-Rey. holds a Doctorate in Science in Electrical Engineering, along with a Master's and Bachelor's in Electronics Systems Design. Since 2014, he has shared his insights as a professor at ITESO University in Mexico, earning the title of Full Professor since 2018.

Speaker 2:

Title: Towards Domain Invariant Block Avoidance: mmWave Recognition Feasibility and OFDM/Radar Dataset

by Bram van Berlo (TU/e M&CS)

Abstract

Joint communication and sensing is envisioned for 6G cellular networks, where operation environment sensing, especially in presence of humans, is as important as the high-speed wireless connectivity. Sensing, and subsequently predicting blockage types in-advance, is an initial step towards signal blockage avoidance. Ideally, advanced data-driven sensing capabilities using deep learning models are unlocked via enough available data over which a distribution portraying a realistic blockage prediction observation space can be interpolated. However, available communication system data volume is limited to prevent design secret disclosure and open-source data simulators provide limited realistic operation environment simulation support. This leads to models that overfit during training and/or suffer from reduced prediction performance when measurement circumstances under which model training data was collected significantly differ from the model inference measurement circumstances. In the presentation, the feasibility of recognizing human motion as a surrogate task for blockage type recognition is discussed with results obtained via visual spectrogram inspection and hyperparameter tuning of deep learning models. A surrogate task can be used for model pre-training, thereby requiring a low amount of data to be collected from the eventual operating environment for fine-tuning. The presentation also introduces a novel mmWave OFDM/Radar dataset containing measurement frames denoting both blockage and human motion scenarios. The frames were captured under a wide variety of different operation environment domains (e.g. walking at different speeds, approaching FoV/boresight at different angles, and different activity start positions). The dataset allows blockage type prediction research under varying domain factors considered during model training and testing.

Bram van Berlo received the EIT Digital dual MSc. degree in embedded systems engineering from the KTH Royal Institute of Technology and the Eindhoven University of Technology (TU/e) with a specialization in entrepreneurship and embedded networking in 2019. He is currently pursuing the Ph.D. degree with the Interconnected Resource-Aware Intelligent Systems (IRIS) Research Cluster, Department of Mathematics and Computer Science, TU/e. His research interests include pervasive sensing and cyber-physical systems. Current research is focused on mitigating domain factor influence among deep learning models used together with unobtrusive radio based (joint communication and) sensing systems.

Webinar call-in details:

________________________________________________________________________________

Click here to join the meeting