Abstract
Radio frequency fingerprint identification (RFFI) uses transmitter-specific hardware imperfections as a physical-layer identity cue for Internet of Things (IoT) devices. However, deep RFFI models often degrade when the acquisition environment changes. In multi-antenna reception, this degradation is not merely a generic distribution shift; it is also shaped by receiver-array topology, frequency-offset dynamics, and capture-dependent target structure, which can distort embeddings and move source-trained decision boundaries.
This article proposes physics-informed structure anchoring with capture-aware prototype calibration (PISA-CAPC), a framework that separates source representation anchoring from fixed-backbone target calibration. The representation stage organizes antenna tokens with a topology graph and modulates the graph using CFO-derived acquisition-dynamics descriptors. Bounded contextual residual suppression is then applied around the identity representation.
At deployment, unlabeled capture-aware prototype calibration (U-CAPC) calibrates target decision scores through capture-local prototype evidence under a fixed representation, mitigating boundary shift without requiring target-domain backbone updates or target labels. On a measured ten-transmitter multi-antenna WiFi benchmark, PISA-CAPC achieves 0.9257 target-domain mean Macro-F1 under a balanced transductive setting. Ablations confirm that topology-guided structure anchoring, contextual residual suppression, and capture-aware calibration contribute complementary gains. These results establish PISA-CAPC as a fixed-backbone route to cross-environment RFFI, coupling physically motivated representation learning with label-free, capture-aware decision calibration.
Blogger's Review: The PISA-CAPC framework effectively addresses the challenges of RF fingerprinting across multiple environments, demonstrating the potential of combining physical insights with deep learning. Its innovative capture-aware calibration strategy significantly enhances model adaptability and accuracy without requiring additional labels, indicating broad applicability in real-world scenarios.