计算机科学
衰退
感知器
继电器
保密
深度学习
人工神经网络
延迟(音频)
卷积神经网络
人工智能
物理层
多层感知器
算法
频道(广播)
无线
计算机网络
电信
功率(物理)
物理
计算机安全
量子力学
作者
Yanyang Zeng,Dawei Zhang,Kai Tang,Jiangfeng Sun
摘要
This paper investigates the physical layer security (PLS) of multi-relay networks over Fisher-Snedecor 𝐹 fading channels. Precise expressions for outage probability (OP), intercept probability (IP), secure outage probability (SOP), and probability of strictly positive secrecy capacity (SPSC) are derived based on this considered model, which the SOP and SPSC are provided in a unified manner using Meijer's G-function. Moreover, simulation consequences of Monte Carlo experiments validate derived mathematical expressions. Through theoretical analysis and simulation results, the interesting results are that larger 𝑚𝑅𝐷, 𝑚𝑠𝑅𝐷, and 𝑚𝑠𝑅𝐸, smaller 𝜇𝑅𝐸, the better secrecy performance of this system. Because deep learning has characteristics of low latency, strong learning ability, and adaptability, in order to pursue lower latency and universal applicability, we combine deep learning and wireless communication. On the basis of constructing a data set with exact closed-form expressions, the secrecy performance is predicted by utilizing a multi-layer perceptron (MLP) deep learning model. Moreover, by comparing the three algorithms of convolutional neural network (CNN), deep neural network (DNN), and two-layer long short-term memory network (LSTM), the MLP has the advantages of higher performance and lower complexity. Experimentations exhibit that the presented algorithm enhances the precision by 75.69% with LSTM for comparison and reduces the time complexity by 80.05% compared with DNN.
科研通智能强力驱动
Strongly Powered by AbleSci AI