Using the theory and method of unascertained measure, an unascertained measure model and the related confidence rule are established to assess the safety state of ship. Thus, the dangerous factors in the hull system c...Using the theory and method of unascertained measure, an unascertained measure model and the related confidence rule are established to assess the safety state of ship. Thus, the dangerous factors in the hull system can be identified, and the accident possibility, loss, and injury degree can be forcasted. An application result shows that the the proposed method is effective in assessment of the traffic safety of ships, and it is more simple in computation than the fuzzy synthetic evaluation method. The proposed method can provide a scientific basis for realizing shipping transportation security and formulating preventive measures.展开更多
舰船网络通信系统的正常运行是保障舰船安全航行的基础。针对现有舰船网络通信系统访问流量异常检测模型检测精度不高和实时性不强的问题,提出一种基于多维度融合注意力的轻量级舰船网络服务器异常流量检测算法。利用Bidirectional Enco...舰船网络通信系统的正常运行是保障舰船安全航行的基础。针对现有舰船网络通信系统访问流量异常检测模型检测精度不高和实时性不强的问题,提出一种基于多维度融合注意力的轻量级舰船网络服务器异常流量检测算法。利用Bidirectional Encoder Representation from Transformers(BERT)作为特征编码器,将捕获的流量数据包映射到深度特征空间;利用深度可分离卷积(Depth-Separable Convolutional, DSC)网络和长短时记忆(Long Short Term Memory, LSTM)神经网络捕获深度编码特征的空间编码特征和时间维度的编码特征;提出一种多维度融合注意力模块,将空间和时间维度的编码特征进行特征融合;利用多维度融合特征进行正常与异常流量的分类。通过在自建的舰船流量异常数据集上进行测试,结果表明所提出模型能够有效检测出舰船网络通信系统的异常访问流量,在保持检测精度的同时,降低了检测时间开销。展开更多
文摘Using the theory and method of unascertained measure, an unascertained measure model and the related confidence rule are established to assess the safety state of ship. Thus, the dangerous factors in the hull system can be identified, and the accident possibility, loss, and injury degree can be forcasted. An application result shows that the the proposed method is effective in assessment of the traffic safety of ships, and it is more simple in computation than the fuzzy synthetic evaluation method. The proposed method can provide a scientific basis for realizing shipping transportation security and formulating preventive measures.
文摘舰船网络通信系统的正常运行是保障舰船安全航行的基础。针对现有舰船网络通信系统访问流量异常检测模型检测精度不高和实时性不强的问题,提出一种基于多维度融合注意力的轻量级舰船网络服务器异常流量检测算法。利用Bidirectional Encoder Representation from Transformers(BERT)作为特征编码器,将捕获的流量数据包映射到深度特征空间;利用深度可分离卷积(Depth-Separable Convolutional, DSC)网络和长短时记忆(Long Short Term Memory, LSTM)神经网络捕获深度编码特征的空间编码特征和时间维度的编码特征;提出一种多维度融合注意力模块,将空间和时间维度的编码特征进行特征融合;利用多维度融合特征进行正常与异常流量的分类。通过在自建的舰船流量异常数据集上进行测试,结果表明所提出模型能够有效检测出舰船网络通信系统的异常访问流量,在保持检测精度的同时,降低了检测时间开销。
