PERFECT: Personalized Federated Learning for CBRS Radar Detection

2026-05-04

AI Summary

This study addresses the critical challenge in dynamic spectrum sharing within the CBRS band: simultaneously protecting naval radar systems from interference caused by commercial LTE/5G transmissions, preserving the privacy of Environmental Sensing Capability (ESC) sensor data, and overcoming detection difficulties arising from non-independent and identically distributed (non-IID) data across geographically dispersed ESC nodes. To this end, the work introduces personalized federated learning into this domain for the first time and proposes the PERFECT framework. In PERFECT, each ESC node trains a local model to ensure data privacy while employing ESC-level personalization to adapt to location-specific radar signal characteristics. Experimental results demonstrate that the proposed approach achieves a 99% recall rate in simulations—meeting regulatory requirements—and matches the performance of centralized solutions while significantly enhancing privacy, computational efficiency, and system scalability.

Abstract

The Citizens Broadband Radio Service (CBRS) band is pivotal for expanding next-generation wireless services, but its success hinges on robustly protecting incumbent users, such as naval radar systems, from interference. This task is delegated to a network of Environmental Sensing Capability (ESC) sensors, which must detect faint radar signals amidst heavy co-channel interference from commercial LTE and 5G users. Traditional centralized detection models raise significant data privacy concerns and are ill-suited for the Non-Independent and Identically Distributed (non-IID) nature of data from geographically dispersed sensors. To overcome these limitations, we propose a novel Federated Learning (FL) framework PERFECT that leverages ESC level personalization for robust and efficient radar detection. PERFECT preserves privacy by training models locally on ESC sensors. Furthermore, our framework is the first to effectively handle non-IID scenarios through model personalization where different ESCs observe distinct radar types. We demonstrate through extensive simulations that PERFECT achieves the mandated 99% recall for radar detection, matching centralized performance while significantly enhancing privacy, efficiency, and scalability for dynamic spectrum sharing.

Problem

• CBRS
• radar detection
• federated learning
• non-IID
• data privacy

Innovation

• Personalized Federated Learning
• CBRS
• Radar Detection
• non-IID
• Environmental Sensing Capability