Q-AudioShield: Quantum-Resilient Audio Watermarking and Authentication Framework for Secure CIoT Multimedia Streams

초록

The proliferation of voice-enabled consumer IoT (CIoT) devices raises critical security concerns regarding audio forgery and privacy breaches, while quantum computing threats render traditional cryptography obsolete. This research presents Q-AudioShield, a quantum-resilient audio watermarking and authentication framework for secure real-time multimedia streams in resource-constrained CIoT environments. The system integrates lattice-based post-quantum cryptography with adaptive spectral watermarking, employing NTRUEncrypt for key exchange and lightweight neural autoencoders for robust watermark detection. Four synergistic components compose the framework: Quantum-Safe Key Management System (QSKMS) with Learning With Errors encryption, Adaptive Spectral Watermarking Engine (ASWE) utilizing psychoacoustic masking, Neural Autoencoder Detection Unit (NADU) with adversarial training, and Integrity Validation Controller (IVC) for threat response. Evaluation across smart speakers, surveillance systems, and mobile assistants demonstrates 15% enhancement in tamper detection accuracy, 0.7-point PESQ improvement, 41.3% latency reduction to 11.1 ms, 21.5% average CPU utilization with 69 MB memory footprint, suitable for resource-constrained devices, and 2.0x quantum security margins (256-bit vs. 128-bit classical baseline, achieving NIST PQC Level 3), establishing new benchmarks for quantum-resilient multimedia security in next-generation IoT ecosystems.

키워드