Research On Wormhole Attacks Detection Methods In Wireless Sensor Networks

Wireless sensor networks constituted by self-organizing nodes use multi-hop wirelesscommunication, and the nodes process information collaboratively. The wormhole attacks canbreak the correct routes in wireless sensor networks. Wormhole consists of two maliciousdistant nodes. The wrong topology information transmission through the wormhole leads tofalse routes.Currently, wormhole attacks can be launched in two modes. One mode is called hiddenmode. The malicious nodes in hidden mode don’t reveal their identities. In addition, onemalicious node captures messages from its neighboring nodes, and sends them through thewormhole to the other malicious node. Then the distant malicious node broadcasts themessages received. The other mode is called participation mode in which two malicious nodesfake the legitimate nodes and are usually far away. But the length of wormhole in theparticipation mode looks like one-hop for other legitimate nodes. Most previous solutionsabout wormhole attacks are dependent on the assumption that the wormhole attacks in acertain mode exists in the networks. Then some measures are proposed based on the certainattacking mode. However, if the attacking mode is different from the assumed mode, then theproposed scheme will not be useful. Because of the uncertainty of the attacking mode, ourapproach is that first make sure whether the network was attacked by the wormhole, thenjudge the corresponding attacking mode. Next, special programs are run to defend against thewormhole attacks according to the judged mode, and secure routes are created finally.This paper proposes a judgement on the attacking mode of wormhole algorithm. Theproposed algorithm can determine whether wormhole attacks happen in the network. If thealgorithm judges that there are wormhole attacks in the network, then it judges the attackingmode of wormhole. This paper defines the abnormal nodes and determines abnormal nodes byfinding abnormal communicatin events. In addition, we analyse the behaviors of wormholeand find that the time when the wormhole attacks happen can not be predetermined. Theresults of experiments show that the wormhole attacks in two modes are both detected with ahigh probability and a low false positive rate. And our method can establish secure routes toavoid the wormhole attacks.