SWIFT-SAT INTERACT: Learning-Based Interference Mitigation for Radiometers

Motivation

Passive microwave radiometers measure faint natural emission to retrieve soil moisture, sea-surface salinity, and other geophysical variables. They share the radio spectrum with a rapidly growing population of active wireless systems, and radio frequency interference (RFI) from those systems corrupts measurements that cannot simply be repeated. Protecting passive science while allowing active users to operate is one of the central spectrum problems of the next decade.

Approach

INTERACT builds an integrated, hardware-in-the-loop testbed in which active wireless transmissions and passive radiometric sensing coexist under controlled and repeatable conditions, and uses it to develop end-to-end learning-based interference mitigation:

• A physical testbed combining software-defined radios, radiometer front ends, and UAS platforms, so that coexistence scenarios can be generated and measured rather than only simulated.
• End-to-end learning architectures that detect, characterize, and mitigate interference directly from raw radiometer data, jointly with the geophysical retrieval task.
• Open datasets released to the community so that other groups can benchmark RFI detection and mitigation methods on realistic, labeled data.

Selected outcomes

• A Physical Testbed and Open Dataset for Passive Sensing and Wireless Communication Spectrum Coexistence, IEEE Access, 2024.
• RFI-Net: Enhancing Passive Sensing through Deep Learning Based Time-Frequency Domain Radio Frequency Interference Detection and Mitigation, IEEE Transactions on Geoscience and Remote Sensing, 2026.