Communication security could be enhanced at physical layer but at the cost of complex algorithms and redundant hardware, which would render traditional physical layer security (PLS) techniques unsuitable for use with resource-constrained communication systems. This work investigates a waveform-defined security (WDS) framework, which differs fundamentally from traditional PLS techniques used in today's systems. The framework is not dependent on channel conditions such as signal power advantage and channel state information (CSI). Therefore, the framework is more reliable than channel dependent beamforming and artificial noise (AN) techniques. In addition, the framework is more than just increasing the cost of eavesdropping. By intentionally tuning waveform patterns to weaken signal feature diversity and enhance feature similarity, eavesdroppers will not be able to identify correctly signal formats. The wrong classification of signal formats would result in subsequent detection errors even when an eavesdropper uses brute-force detection techniques. To get a robust WDS framework, three impact factors, namely training data feature, oversampling factor and bandwidth compression factor (BCF) offset, are investigated. An optimal WDS waveform pattern is obtained at the end after a joint study of the three factors. To ensure a valid eavesdropping model, artificial intelligence (AI) dependent signal classifiers are designed followed by optimal performance achievable signal detectors. To show the compatibility in available communication systems, the WDS framework is successfully integrated in IEEE 802.11a with nearly no adding computational complexity. Finally, a low-cost software-defined radio (SDR) experiment is designed to verify the feasibility of the WDS framework in resource-constrained communications.
翻译:可以用复杂的算法和冗余硬件来增强物理层的通信安全,但这会使传统的物理层安全(PLS)技术不适于使用资源限制的通信系统。这项工作调查了波形定义的安全(WDS)框架,它与当今系统使用的传统PLS技术有根本的不同。框架并不取决于频道条件,例如信号功率优势和频道状态信息(CSI),因此,框架比频道依赖光束和人工噪音(AN)技术更可靠。此外,框架不仅仅是增加窃听的低成本。通过有意调整波形格式模式以削弱信号特性多样性和增强特征相似性,eavesdropers将无法正确识别信号格式。对信号格式的错误分类将会导致随后的检测错误,即使电子播音器使用布鲁氏探测技术。要获得一个强有力的WDSDS框架,三个影响因素,即培训数据特性、过度扫描系数和带宽压缩系数(BCFCF)抵消。通过有意调低的网络格式调整,将一个最佳的WDS-S-road Forld Streal 格式框架与一个在联合研究后,最后将显示一个最佳的SAL-deal-laeval-de Axxx 。