Title: Ku-Band and C-band Enhanced MetaSurface Radar with CMOS System-on-Chip Using Hardware Machine Learning for Remote Sensing of Snow: Winter 2023 and 2024 Field Experiments
Presenting Author: Hans-Peter Marshall
Organization: Cryosphere Geophysics and Remote Sensing
Co-Author(s): Adrian Tang, Nacer Chahat, Thomas Van Der Weide, Yanghyo Kim, Goutam Chattopadhyay, M-C Frank Chang

Abstract:
Remote sensing of seasonal snow mass, which represents 70% of water supply in the Western U.S., remains challenging, with no operational satellite-based global approach. Airborne and field experiments are needed to better understand scattering mechanisms as a function of frequency, for both time-of-flight/phase-based approaches, as well as volume-scattering amplitude-based techniques. These novel radar systems have been integrated with a heavy-lift UAV, allowing frequent deployments, at much lower cost and logistical effort compared to crewed aircraft deployments. This enables a time series approach, throughout the full winter season, expanding the dynamic range of conditions. The radar system is implemented with a set of custom CMOS chips that provide all the required radar functions: waveform generation, upconversion, downconversion and radar waveform correlation / demodulation. The radar chipset also contains an embedded Machine Learning engine which dynamically adjusts the radar waveforms to minimize the transmitter-to-receiver leakage in the radar, leading to enhanced isolation, higher dynamic range, and more sensitivity to the snowpack. This radar chip is coupled with a compact, lightweight, and high-gain all-metal metasurface antenna which also offers intrinsically high isolation between transmit and receiver ports. This presentation reviews the chipset and antenna design, and presents results from the past 2 winters of field experiments.