Dual-Security for Indoor OFDM-ISAC Systems via Temporal Artificial Noise

Technical Summary: Dual-Security for Indoor OFDM-ISAC Systems via Temporal Artificial Noise

Problem Statement
Integrated Sensing and Communication (ISAC) systems, particularly those based on Orthogonal Frequency Division Multiplexing (OFDM) in indoor environments, face a unique dual-security challenge. Unlike traditional scenarios where threats are external, this paper addresses a "pathological" case involving authorized users who act as malicious eavesdroppers for services they are not permitted to access. Specifically, the system considers:

1. Sensing Users (SU): Authorized for sensing tasks (e.g., target impulse response estimation) but unauthorized for communication data. They may attempt to eavesdrop on communication payloads.
2. Communication Users (CU): Authorized for data transmission but unauthorized for sensing. They may attempt to perform unauthorized sensing on targets.

Existing literature typically addresses either communication security (data privacy) or sensing security (target privacy) in isolation. Furthermore, indoor ISAC poses specific risks because Channel State Information (CSI) used for sensing can reveal personal identifiable information (e.g., vital signs), violating regulations like the EU GDPR. Current secure indoor sensing designs often fail to guarantee communication performance, rendering them unsuitable for true ISAC systems.

Methodology
The authors propose a Temporal Artificial Noise (AN) scheme embedded within the Cyclic Prefix (CP) of the OFDM waveform to achieve dual-security. The system model assumes a single-antenna transmitter, one CU, one SU, and one target, with quasi-static channels and temporally separated communication and echo signals.

The core mechanism relies on the duality between time and frequency domains and the specific processing steps of OFDM receivers:

• AN Design: The AN vector is superimposed on the signal and designed to lie in the null space of the authorized Communication User's channel ($H_{CU}$).
• Communication Security (Protecting Data from SU): When the SU attempts to decode communication data, it removes the CP and applies an FFT. Because the AN is designed based on $H_{CU}$, it does not fall into the null space of the SU's channel ($H_{SU}$). Consequently, the SU receives the AN as interference in the frequency domain, degrading its Signal-to-Noise-plus-Interference Ratio (SINR) and preventing successful data decoding.
• Sensing Security (Protecting Targets from CU): Both the SU and CU receive the target echo signal.
  • The SU acts as a cooperative bistatic receiver, using the full transmitted frame (including the AN in the CP) as a reference signal. Since the AN is present in both the reference and the echo, the SU can accurately estimate the target channel, maintaining high sensing performance.
  • The CU, attempting unauthorized sensing, lacks knowledge of the AN. It treats the AN as part of the noise or interference when estimating the target impulse response. This creates a mismatch between the CU's reference signal (data only) and the received echo (data + AN), significantly increasing the Mean Squared Error (MSE) of the channel estimation and degrading sensing performance.

Key Contributions

1. Dual-Threat Model: The paper establishes a security framework where authorized users are treated as potential adversaries for unauthorized services, a scenario often overlooked in favor of external eavesdroppers.
2. Temporal AN Design: A novel scheme is proposed that embeds AN in the CP. This design exploits the fact that standard OFDM receivers discard the CP for communication (removing the AN for the CU) but require the full signal structure for sensing (retaining the AN for the SU).
3. Optimization Framework: The authors formulate a convex optimization problem to maximize legitimate service quality (CU communication SNR and SU sensing accuracy) while satisfying security constraints (SU decoding SINR and CU sensing MSE). The problem is transformed using Schur complements and trace-based relaxations to ensure global optimality.
4. Performance Guarantees: The scheme guarantees the Quality of Service (QoS) for legitimate tasks while simultaneously suppressing unauthorized activities without requiring additional hardware or complex waveform modifications beyond the CP.

Simulation Results
Numerical results demonstrate the effectiveness of the proposed design under various power budgets and threshold constraints:

• Communication Security: As the requirement for communication security tightens (lower SU decoding SINR threshold), the system allocates more power to AN. This reduces the CU's communication SNR slightly but effectively suppresses the SU's ability to decode data.
• Sensing Security: Increasing the power budget improves the sensing performance gap between the SU and CU. The SU maintains low estimation error regardless of AN power allocation, while the CU's estimation error increases significantly when AN is present, ensuring sensing security.
• Trade-offs: The results show that enhancing both communication and sensing security requires increased power allocation to the AN. While this slightly degrades the CU's communication performance, the SU's sensing capability remains largely unaffected, confirming the dual-protection capability.

Significance
The paper claims to bridge a critical gap in ISAC research by addressing dual-security (protecting both communication and sensing functions simultaneously) in indoor environments. By leveraging the structural properties of OFDM (specifically the CP) and temporal AN, the proposed method provides a physical layer security solution that:

• Prevents authorized users from eavesdropping on unauthorized services.
• Ensures that legitimate sensing and communication performances are preserved.
• Offers a practical approach for indoor ISAC systems where CSI-based sensing poses privacy risks, aligning with data protection regulations.

The work suggests that temporal AN is a viable mechanism to secure ISAC systems against internal threats without compromising the utility of the shared spectrum and hardware resources.