Research On Intrusion Detection Technology For Wireless Sensor Networks

Wireless sensor networks(WSN), which intergrate the technologies of sensor, micro-electro-mechanism system, wireless communication,and distributed computing, are the next generation emerging network.The applications of wireless sensor networks in military, medical treatment and environmental protection, etc became more and more, But the following security problem also becomes a big barrier of widespread WSN application. Though a variety of traditonal security mechanisms, such as firewall, authentification,and data encryption, are applied in wireless networks widely, they have many defects. The intrusion detection technology, as a supplement, can make up for deficiency of intrusion prevention approaches,which is one of the hotspot problems in WSN.Because WSN node constraints by various resource, such as energy consumption restriction, computing capablility, In the case of various intrusion behavior, it is easy to be invading main target. In this paper, current WSN intrusion detection technology analyzed firstly. The status of WSN faces intrusion behavior and characteristics are surveyed in details. Then, paper reviews the two major problem, which is bad Instantaneity and low efficiency caused by energy consumption restriction, in WSN instrusion detection system. According to the content of the above, double clumping wireless sensor network intrusion detection system structure is proposed in this paper. We make use of lightweight Mahalanobis Distance to identify attackers in the cluster. This methods collect multiple network attributes. Multiple attributes of the sensor nodes are taken into consideration to find the pernicious attacker in the cluster.Due to the limited energy of sensor node, the instrusion detection in the cluster use a lightweight algorithm. In order to improve accuracy of the whole instrusion detection system, this system include the Sink node. A complex intrusion detection algorithm is used in the Sink node since it has enough resources to make up of double clumps wireless sensor network intrusion detection system together with cluster detection. Because of various upper application, this paper proposed the detection node selection algorithm based on the remaining energy to improve energy consumption of detection node.This paper use TOSSIM + Graphviz simulation platform as experiment entironment after system design and combining professional CUP99 data set which is a base data of simulation aggressive behavior to make a comparison between our algorithm and other intrusion detection algorithm. The experiment shows that our system can achieve highter detection accracy rate and get a lower false alarm rate.Finally, this paper summarizes the research achievements and illuminates the future research direction.