多传感器网络信息数据融合技术研究

Research of the Data Fusion Technique for Multisensor Network Information

本文提出了多传感器网络信息数据融合技术。通过多传感器探测节点向目标主机发送探测信息,并将返回信 息输入到经过大量产生式规则训练和学习的BP神经网络中,得到目标主机属性信息。然后,通过D-S证据推理对属性信 息进行信任度融合处理,得到目标主机上可能存在的漏洞分布及漏洞存在的可能性,为进一步的主机安全性能评估打下基 础。

This paper puts forward the fusion technique of multisensors network information. The sensors send data to the target host, and the returned information from the target host will be inputted to the BP neural network, which has undergone training and learning by a lot of production rules, and then the attribute information of the host will be acquired. Afterwards, these attribute information will be fused by the D-S evidence theory. As a result, the existence probability of vulnerabilities and the distribution of each vulnerability are presented, which paves the way for security performance evaluation of the host