Abstract The spectrum sensing scheme suffers from a physical layer attack of Primary User Emulation Attack (PUEA). The resolution is to mitigate the cognitive radio user from the PUEA under the physical layer. Detecting the PUEA attack in real-time is a challenging one. The traditional Location-based PUEA detection requires the primary user's location knowledge, which may not be possible practically. This research focuses on developing a reliable spectrum sensing mechanism in the presence of PUEA attack and rapid change in the wireless channel. This reliable spectrum sensing framework is developed using the neural network-based PUEA detector excluding the location information. The Software-Defined Radio (SDR) called Universal Software Radio Peripheral (USRP) 2943R is used to implement the proposed mechanism for analyzing performance in real-time. The real-time experimental results show that PUEA detection can be achieved with 97% accuracy.

中文翻译：

---

使用神经网络缓解主要用户模拟攻击

摘要 频谱感知方案受到主用户仿真攻击（PUEA）的物理层攻击。解决方案是从物理层下的 PUEA 中减轻认知无线电用户的影响。实时检测 PUEA 攻击是一项具有挑战性的工作。传统的基于位置的 PUEA 检测需要主要用户的位置知识，这在实践中可能是不可能的。这项研究的重点是在存在 PUEA 攻击和无线信道快速变化的情况下开发可靠的频谱感知机制。这种可靠的频谱感知框架是使用基于神经网络的 PUEA 检测器开发的，不包括位置信息。称为通用软件无线电外设 (USRP) 2943R 的软件定义无线电 (SDR) 用于实现建议的实时性能分析机制。实时实验结果表明PUEA检测可以达到97%的准确率。