Physical Layer Security (PLS)

The process of uniquely identifying a specific physical radio transmitter by analyzing the distinct, unintentional features embedded in its emitted waveform. SEI operates independently of the encoded data or modulation scheme, relying instead on hardware-intrinsic impairments such as oscillator drift and amplifier non-linearity. This technique is foundational for physical layer authentication, enabling a receiver to distinguish between two otherwise identical radio models.

• Extracts features from transient turn-on signatures and steady-state emissions
• Used in military SIGINT and critical infrastructure access control
• Complements cryptographic methods by binding identity to physical hardware

A hardware security primitive that derives a unique, unclonable cryptographic key from the inherent, random physical variations introduced during semiconductor manufacturing. A PUF implements a Challenge-Response Pair (CRP) mechanism where a digital input stimulus produces a deterministic, physically-derived output unique to that specific silicon instance. Unlike stored keys, PUF-derived secrets are not present in memory when the device is powered off, providing strong resistance to physical intrusion.

• Exploits process variation in transistor gate delays or SRAM power-up states
• Serves as a root of trust for key generation and device authentication
• Eliminates the need for secure non-volatile key storage on IoT devices

A biometric-like profile of a wireless device constructed from the aggregate of its unique, hardware-intrinsic signal imperfections. RF-DNA combines multiple discriminative features—including phase noise fingerprints, I/Q imbalance, and power amplifier non-linearity—into a composite identity vector. This profile remains consistent across different transmitted data payloads, enabling passive identification without requiring cooperation from the target device.

• Leverages bispectrum analysis to capture non-linear phase couplings
• Requires drift compensation models to account for temperature and aging effects
• Evaluated using Equal Error Rate (EER) as the primary authentication metric

The detailed amplitude and phase information describing how a radio signal propagates from a transmitter to a receiver. In PLS, CSI serves as a location-bound, environment-dependent feature for authenticating a device's physical position. Because the multipath channel is reciprocal and decorrelates rapidly over half a wavelength, an attacker at a different location cannot replicate the legitimate user's CSI, providing inherent replay attack resistance.

• Enables distance bounding protocols to defeat relay attacks
• Used in continuous authentication schemes for session-long verification
• Highly dynamic; requires frequent re-estimation in mobile environments

A classification paradigm where a model must accurately identify known, enrolled devices while simultaneously detecting and rejecting any previously unseen or rogue emitters. In the context of PLS, an open set classifier assigns an 'unknown' label to signals from devices not present in the training gallery. This capability is critical for real-world deployment, where the system must gracefully handle novel hardware and adversarial impersonation attempts.

• Uses contrastive learning to create discriminative embedding spaces
• Requires careful threshold calibration on Equal Error Rate (EER) curves
• Essential for detecting hardware trojans and counterfeit components in supply chains

A privacy-preserving machine learning framework where multiple distributed receivers collaboratively train a shared device identification model without exchanging raw signal data. Each receiver computes local model updates on its own captured emissions and shares only the gradient updates with a central aggregation server. This architecture is vital for sovereign AI infrastructure and defense applications where raw IQ samples cannot leave a secure facility.

• Protects sensitive unintentional radiated emissions data from exposure
• Enables collaborative model improvement across geographically dispersed SIGINT sites
• Requires robust aggregation protocols to handle non-IID data distributions across receivers