中文摘要：

从节点行为对网络安全的影响角度看,恶意节点将直接导致路由破坏或者毁坏数据.因此,在节点行为检测和信任度评估时,必须首先重点关注其安全性行为特征,以便降低恶意行为节点的信任值.归纳了网络中九种常见的节点行为类型,分析了各行为状态的行为特征,建立了特征模型,提出了检测特征参数和关联检测模型.仿真结果表明,提出的节点行为检测特征参数和关联检测模型,在静态及网络变化时,对不合作节点、恶意节点的行为信任值评估定级均有较高的准确性,可以有效排除这些节点的网络活动.

英文摘要：

In the angle of networks security effects, malicious node behaviors lead to route failure and data destroy. So node behavior security evaluation is the most important evaluation factor in behavior detection and trust evaluation, which can debase trust value of the malicious nodes. This paper has an analysis on 9 sorts of common node behaviors profile, which presents behavior detection parameters and detection model. The results show that in the static and dynamic network environment it reaches a high detection accuracy and evaluation accuracy which can exclude malicious node from network activity effectively.