﻿ 基于云计算的舰船通信网络入侵检测方法研究
 舰船科学技术  2024, Vol. 46 Issue (10): 170-173    DOI: 10.3404/j.issn.1672-7649.2024.10.030 PDF

Research on intrusion detection methods for ship communication networks based on cloud computing
HUANG Guo-feng, LIU Yu-ping
Wuhan Institute of Shipbuilding Technology, Wuhan 430050, China
Abstract: To improve the real-time performance of intrusion detection in the face of large-scale network attacks or sudden attacks, a cloud computing based intrusion detection method for ship communication networks is studied. Using the MapReduce programming model of cloud computing, design a genetic quantum particle swarm optimization algorithm for MapReduce parallelization, and extract network intrusion features from ship communication network data; Using the MapReduce parallelized entropy clustering algorithm, determine the basis function center of the radial basis function neural network; After determining the center of the basis function, input network intrusion feature samples into the Map function of the MapReduce programming model, train the neural network, optimize the weights of the neural network, output training completion instructions through the Reduce function, and complete the neural network training; In the trained MapReduce parallelized radial basis function neural network, input feature samples and output intrusion detection results for ship communication networks. Experimental results have shown that this method can effectively extract intrusion features from ship communication networks; This method can accurately detect ship communication network intrusion under different types of network attacks.
Key words: cloud computing     ship communication network     intrusion detection     mapreduce parallelism     particle swarm     radial basis function
0 引　言

1 舰船通信网络入侵检测方法 1.1 基于云计算的舰船通信网络入侵特征提取

 $S\left( {A,B} \right) = \frac{{2\left[ {E\left( A \right) + E\left( B \right) - E\left( {A,B} \right)} \right]}}{{E\left( A \right) + E\left( B \right)}} 。$ (1)

f的计算公式如下：

 $f = \frac{{\sum\limits_l {S\left( {{A_l},C} \right)} }}{{\sum\limits_l {S\left( {{A_l},{B_l}} \right)} }}。$ (2)

1.2 基于云计算的网络入侵检测

 ${\varphi _{\hat i}} = {e^{ - \frac{{{{\left\| {{z_{\hat j}} - {\theta _{\hat i}}} \right\|}^2}}}{{2\delta _{\hat i}^2}}}}。$ (6)

 ${y_{\hat j}} = \sum\limits_{\hat j = 1}^\eta {\sum\limits_{\hat i = 1}^\rho {{w_{\hat i\hat j}}{\varphi _{\hat i}}} }。$ (7)

RBF神经网络中，基函数中心直接影响舰船通信网络入侵检测精度，为此，通过云计算的MapReduce编程模型，设计MapReduce并行化熵聚类算法，优化RBF神经网络的基函数中心，具体为：

1）在Map函数内，输入提取的舰船通信网络入侵特征，计算侵特征样本和间${\theta _{\hat i}}$的熵值${H_{\hat j}}$

2）按照最小${H_{\hat j}}$，归类各舰船通信网络入侵特征样本，输出聚类结果。

3）利用Reduce函数内的Combine函数，求和处理聚类结果，确定更新后的基函数中心。

2 结果分析

 图 1 部分舰船通信网络拓扑结构图 Fig. 1 Partial topology diagram of ship communication network

 图 2 舰船通信网络入侵检测结果 Fig. 2 Intrusion detection results of ship communication network

 图 3 MCC分析结果 Fig. 3 MCC analysis results
3 结　语

 [1] 陈立家, 周为, 许毅, 等. 一种基于SDN的多约束无人船网络传输路由算法[J]. 中国舰船研究, 2022, 17(4): 107-113. [2] 尹晟霖, 张兴兰, 左利宇. 双重路由深层胶囊网络的入侵检测系统[J]. 计算机研究与发展, 2022, 59(2): 418-429. [3] 马明艳, 陈伟, 吴礼发. 基于CNN_BiLSTM网络的入侵检测方法[J]. 计算机工程与应用, 2022, 58(10): 116-124. [4] 景雯, 张杰. 基于区块链技术的无线传感网络入侵检测算法[J]. 传感技术学报, 2023, 36(6): 978-983. [5] 刘拥民, 杨钰津, 罗皓懿, 等. 基于双向循环生成对抗网络的无线传感网入侵检测方法[J]. 计算机应用, 2023, 43(1): 160-168. [6] 唐玺博, 张立民, 钟兆根. 基于ADASYN与改进残差网络的入侵流量检测识别[J]. 系统工程与电子技术, 2022, 44(12): 3850-3862.