Tu1G: AI for Advanced Radar and Modulation Sensing

Back to IMS Schedule
Tue 9 Jun | 08:00 - 09:40
157AB
Akim Babenko
Google
Chris Thomas
Boeing
This session highlights the integration of Machine Learning and Digital Signal Processing to solve challenges in the RF and mm-wave domains. The presentations highlight innovations such as Spiking Neural Networks on FPGAs for high-speed modulation recognition, PointMLP architectures for sparse radar data classification, physics-informed state-space models for robust tracking in multipath environments, and multimodal IR-Radar fusion to ensure privacy-preserving event recognition.
08:00 - 08:20
Tu1G-1 Real-Time, Over-the-Air Modulation Recognition using a 320 MS/s Spiking Neural Network on FPGA
ABSTRACT
Sai Sanjeet, Bibhu Datta Sahoo
Univ. at Buffalo, Univ. at Buffalo
08:20 - 08:40
Tu1G-2 Sparsity-Based Range-Velocity-Time PointMLP for FMCW Radar Human Activity Classification with Efficient Computation
ABSTRACT
Yu-Hong Wu, Chin-Lung Yang
National Cheng Kung Univ., National Cheng Kung Univ.
08:40 - 09:00
Tu1G-3 Parallel Kalman Filtering with Physics-Informed Selective State-Space Models for Robust Radar Sensing
ABSTRACT
Jinhyeok Park, Apurba Padhy, Han Cho, Saibal Mukhopadhyay
Georgia Institute of Technology, Georgia Institute of Technology, Georgia Institute of Technology, Georgia Institute of Technology
09:00 - 09:20
Tu1G-4 Learning to Track: Deep Association for Multipath-Resilient In-Air Writing with D-band FMCW Radar
ABSTRACT
Salah Abouzaid, Leander Nothelle, Nils Pohl
Ruhr Univ. Bochum, Ruhr Univ. Bochum, Ruhr Univ. Bochum
09:20 - 09:40
Tu1G-5 Millimeter-wave 3D Radar and IR Multimodal Sensing System Enabling AI-based Event Recognition with Enhanced Privacy
ABSTRACT
Asaf Tzadok, Stanislav Lukashov, Alberto Valdes-Garcia
IBM T.J. Watson Research Center, IBM Research, IBM T.J. Watson Research Center