Gesture Recognition from Radar Simulations

Radar-based gesture recognition algorithms can work without a single measured sample for training! In a new journal article by Nicolai Kern, this is demonstrated by one-to-one re-simulation of a comprehensive real dataset, thus pointing out both the potential and the challenges of simulation-based data augmentation. The results are presented in the Transactions on Microwave Theories & Techniques’ special issue on AI for Microwaves, now available in Early Access.

Paper: Learning on Multistatic Simulation Data for Radar-Based Automotive Gesture Recognition

EuRAD Young Engineer Prize for Pirmin Schöder

We congratulate Pirmin Schöder on receiving the EuRAD Young Engineer Prize for his paper “A Modulation-Based Target Simulator and Its Hardware Nonidealities”.

Taking a Look Beneath the Surface

Multicopter UAVs offer unmatched flexibility, which allows for new ways for close-proximity ground-penetrating radar imaging without requiring traversable terrain. There is a large variety of applications that could benefit from these systems. In a new magazine article, Alexander Grathwohl and Michael Stelzig provide an overview of the requirements for high-quality imaging, current state-of-the-art applications and future possibilities. The system prototype is even featured on the IEEE Microwave Magazine's cover.

Paper: Taking a Look Beneath the Surface: Multicopter UAV-Based Ground-Penetrating Imaging Radars

European Microwave Week in Milano

We are contributing seven papers and two workshops presentations to the conference taking place end of September.

MIMO Radar with 1728 Channels

A new MIMO Radar with 1728 Virtual Channels for high-resolution 4D imaging is presented by Dominik Schwarz in his new journal article. The system is composed from 12 radar MMICs, exceeding the channel count of modern automotive radars by a factor of 10, has a new multilayer PCB concept to seperate RF circuitry and antennas, and features beamwidths of 0.78° (azimuth) and 3.6° (elevation). The article is already available on IEEExplore and will be published in the Journal of Microwaves.

Paper: System Performance of a 79 GHz High-Resolution 4D Imaging MIMO Radar With 1728 Virtual Channels

Int. Radar Symposium

The institute contributes a paper about Interference of Automotive Radars in Complex Traffic Scenarios to the conference.

Radar-Based Gesture Recognition with Neural Networks

Neural networks can reliably recognize real-world traffic gestures from multi-perspective radar target lists as input. Hence, the gesture recognition algorithm presented by Nicolai Kern in his article for the Transactions on Aerospace and Electronic Systems is particularly suitable for autonomous driving in strongly scattering environments, helping vehicles to safely interact with pedestrians. The article is already available on IEEExplore Early Access.

Paper: PointNet+LSTM for Target List-based Gesture Recognition with Incoherent Radar Networks

Prize for Doctorate

Martin Geiger was awarded a prize for his outstanding dissertation, donated by the Ulmer Universitätsgesellschaft. Congratulations! Martin wrote his dissertation about a flexible waveguide probe for MMIC-based radar systems at 300 GHz.