This research paper introduces a novel low-complexity physical layer security technique based on an encryption model for fixed LiFi communication systems. This encryption model is based on the Channel Impulse Response (CIR) between the two users. This technique shows to be efficient against eavesdropping and self-aware of malicious activities. The attacker will get a Bit Error Ratio (BER) of over 0.4 which will disable him from deciphering the transmitted signal and the same thing happens when a user receives malicious information. Also, it helps construct information on the CIR of a potential attacker using machine learning techniques, with an accuracy of over 0.85. Such information can be efficient in case the attacker is a user within the network, i.e. its CIR is known by the receiver, as it helps revealing his identity and ultimately blocking him from the system.

本研究论文介绍了一种基于固定LiFi通信系统加密模型的新型低复杂度物理层安全技术。此加密模型基于两个用户之间的信道冲激响应（CIR）。该技术显示了对窃听和自我感知恶意活动的效率。攻击者将获得超过0.4的误码率（BER），这将使他无法解密传输的信号，并且当用户收到恶意信息时也会发生同样的事情。此外，它还使用机器学习技术帮助构建有关潜在攻击者CIR的信息，准确性超过0.85。如果攻击者是网络内的用户，即接收者知道其CIR，则此类信息可能非常有效。