An attack identifi cation technique based on behaviour analisys for wireless sensor network

Korzhuk Viktoriia Mikhailovna

This article presents a technique for identifying attacks on wireless sensor networks. The developed technique allows using a certain set of features to identify 14 types of network-level attacks. The article describes and justifies the use of computer programs developed during the study, which are necessary to increase the efficiency of identification of attacks on wireless sensor networks. The random forest algorithm and the probabilistic classifier are used together for identification. Experiments on mesh and cluster topology networks were carried out. Recommendations on the application of the technique are offered. The results can be used as part of intrusion detection systems to improve the availability and integrity of information.

Keywords:cyber-physical systems, cyber security, information security, wireless sensor networks, attack identification, identification methodic