Research On Trust-based Intrusion Detection In Wireless Sensor Networks

The development of micro-chip and short distance communication technique drives Wireless Sensor Networks(WSNs)to be applied into various scenarios.Due to the limitations of resource and unattended deploying manner,WSNs are faced more serious security challenges than the cable network.Traditional security techniques,such as encryption and authentication,can protect WSNs from being eavesdropped or tampered.However,deploying these technologies will greatly affect the performance of sensor due to the limitations of the sensor resources.Therefore,it is necessary to exploit new attack detection and defense techniques for WSNs.In this paper,the trust-based intrusion detection technology in a WSN is taken as the research object,and the three schemes including the network topology,the intrusion detection scheme and the data aggregation method are studied.The performance of WNS is improved from the aspects of energy consumption,security and aggregation accuracy.Firstly,we construct a network topology suitable for intrusion detection technology.Because the existing trust scheme ignores the topology construction of WSNs,many challenging issues need to be conquered in the energy control.Therefore,this paper proposes a network topology which is more suitable for intrusion detection mechanism under the premise of energy consumption and topological stability.The communication distance,the residual energy and the node density are used as references for establishing connections between nodes that make the connections between nodes more reasonable and improve the stability of network topology.Secondly,an intrusion detection mechanism based on trust is discussed.As an active defense technology,intrusion detection can not only identify potential attacks,but also protect the monitored network from being compromised through deploying the security countermeasures.Meanwhile,it provides relevant data records that can be used as a data source to track the malicious activity.Therefore,we introduce the intrusion detection mechanism into WSNs and propose a Beta Trust-based Intrusion Detection System(TIDS)to evaluate the reliability of the node.In addition,the existing model fails to effectively evaluate the trustworthiness of the node from multiple aspects and there are deficiencies in the security and applicability of the network.Therefore,a comprehensive trust evaluation model based on multi-index is proposed.The direct trust of node is calculated according to the energy consumption,the data accuracy and the forwarding behavior of nodes.The direct trust is a combination of data trust,energy trust and forwarding trust value.The indirect trust is evaluated according to the communication behavior and the recommended credibility of the neighbor nodes.Therefore,the total trust value of node is calculated according to the direct and indirect trust value.Finally,the trusted next-hop node isselected according to the self-recommended trust value of the node(the willingness of participating in the communication behavior between nodes),so that the data aggregation can be completed efficiently and accurately.Finally,a security aggregation approach is proposed.According to the network scale and hierarchy,the sensor node,the cluster head and the base station are regarded as the basic elements and a trust aggregation model based on multi-level cluster-based hybrid topology is designed.Taking the security and accuracy of data aggregation into consideration,this model processes the perceived data from part to whole which can effectively obtain the trusted data.After that,the data are uploaded to upstream node hop by hop according to the actual application requirements.The processes of data aggregation can provide the security data aggregation,promote the data aggregation accuracy,reduce the amount of data transferred to the upper layer and improve the utilization of node energy effectively.