Wireless Sensor Network Sinkhole Attack Detection Algorithm Based On Reputation Evaluation

Wireless Sensors Network(WSN)is a kind of self-organized,multi-hop,energy and storage constrained network composed of multiple sensing nodes.WSNs are widely used in many aspects,due to the limitations of communication radius,battery capacity,storage and computation,it is very easy to be damaged and become a failure node in the external environment,which poses a serious threat to the connectivity and data accuracy of the network,and poses a great challenge to the security of the network.WSNs are generally vulnerable to two different types of attacks,internal and external.External attack is a kind of attack that can obtain the detailed information of the internal nodes of the network without internal authentication,which can be resisted by encryption authentication technology,but the technology cannot resist internal attacks,internal attacks by capturing the internal nodes of the network and stealing internal information for malicious operations,including sinkhole attacks,wormhole attacks and selective forwarding attacks and other internal attacks,attacks will lead to poor connectivity of the network,the data low accuracy and affect the security of the network,because of the difficulty of implementing detection and defense mechanisms,it is very important to identify and isolate the malicious nodes.This paper introduces WSNs from the basic concepts,composition structure,network characteristics and security-related theories,etc.It also deeply researches the Sinkhole attack detection algorithm and systematically describes about WSN security-related principles,including WSN attack types and typical reputation models.This paper aims to solve the problem of WSN network security reduction due to Sinkhole attack,and to research the relevant security detection technologies in wireless sensing networks.Firstly,a dynamic trust evaluation model integrating fuzzy comprehensive evaluation mechanism and similarity measure theory is proposed,abbreviated FSEPM.Based on the trust model,the state of nodes can be judged by the interaction behavior between nodes,and the malicious nodes within the network can be identified timely and accurately,and the internal misbehavior can be discovered and eliminated.However,there are problems of subjectivity and fuzziness of inter-node trust,speed and accuracy of trust model for attack detection.In this paper,we use fuzzy comprehensive evaluation,similarity measure theory and weighting factor to improve the problems caused by subjective assignment of weights,combined with AR model to establish the model of predicted value of trust value,and compare the difference calculation between predicted and actual trust value with trust threshold to judge the nature of nodes,detect malicious nodes,secure the security of wireless sensor network and achieve reliable transmission of data.For the Sinkhole attack which is difficult to detect and defend,a Sinkhole attack detection and defense strategy integrating SPA and Jaya algorithm in WSNs is proposed,abbreviated as SJ-SHDDS.The boundary nodes with abnormal hops are detected by the anomaly degree,and then the trust values of suspicious nodes are calculated by combining with the SPA trust model which can deal with the interaction of certainty and uncertainty,and the attack nodes are detected;the Jaya algorithm that has the advantages of parameter-free operation,fast solution speed,and not easy to fall into the local optimal solution is used to avoid the attacked area to send the attack node information to the real Sink.Simulation results show that the improved SJ-SHDDS can effectively detect and defend against malicious nodes and improve the security of wireless sensor networks.