Research Of Situation Awareness Mechanism Based On Trust In Wireless Networks

With the appearances of ubiquitous network and convergence technology, there area sharp increasing of network data and frequent communications, which is a challengefor researchers to realize network security and adaptation. Owing to wireless sensornetworks with limited resources and capabilities joining in network fusion, the wirelessnetworks appear the characteristics of vulnerability and dynamic change, which will beconvenient for adversary to falsify and fabricate network entity data. Therefore, thetechnology of network security situation awareness arises. In this area, it has been aneffective way of building situation awareness mechanism based on trust to resistmisbehaviors, thus can sense the misbehaviors. That can also make decisions accordingto dynamic situation, and provide advantages for network management and working.This thesis researches situation awareness mechanism based on trust in wirelessnetwork, mainly studying the follow aspects:Firstly, attacks are always seeking ways of exploiting any existing weakness inwireless network. Aiming at this problem, regarding social network as the researchobject, we propose a decision-aided situation awareness mechanism based onmulti-scale dynamic trust from the perspectives of time and space, which can recognizeand handle misbehaviors. We build trust based on Ebbinghaus forgetting regular andspatial correlations. The mechanism carries out decision making through trustauthenticity test, logicality test and feedback parameters. Simulation analysisdemonstrates that compared with other trust mechanisms, this mechanism can recognizeand handle misbehaviours more effectively, which is relatively eclectic and realistic fortrust mensuration. Additionally, it can effectively avoid resource congestion.Secondly, to cope with the deviation problem caused by captured nodes in wirelesssensor network, we present a situation data fusion mechanism based on trust. It makestrust awareness rule based on historical trust or data correlation, and processes data toguarantee the integrality in three stages: At first, in event detection, after collecting data,it detects events and pretreats data through most trust majority rule, to enhance the dataaccuracy. Then, during data fusion, it uses data filter rule to improve data reliability. Itallocates data weights depending on trusts. Finally, in consistency detection, it utilizesthe consistency detection rule to reduce communication traffic, simultaneously judging whether the centers are credible or not. The new mechanism can suppressmisrepresentation and enhance the performance when abnormities are more than normalones. It also can reduce the deviation between real value and fusion value. Theoryanalysis and simulation results have verified the reliability and validity.