Development of a new hybrid intrusion detection system in wireless sensor networks

Wireless sensor networks (WSNs) are used in many areas today to detect variables such as temperature, pressure, vibration, speed in physical environments and to produce results. These networks are also used in critical and vital areas such as military, intelligence, health and natural disasters, where data security is important. However, since WSNs have different infrastructure features and hardware constraints than traditional networks, it is essential to take effective security measures for these networks. With the developing technology and changing conditions, security measures have been developed for WSNs. One of the effective security mechanisms recommended for WSNs is intrusion detection system (IDS). However, in today's conditions, the detection methods recommended for IDSs are not sufficient on their own to ensure security and the use of outdated data sets in most of the existing studies has necessitated an effective and up-to-date IDS model for WSN security. In this thesis, a model has been developed in which the proposed detection methods for WSNs are used as hybrids. In the model, due to resource constraints of WSNs, preprocessing steps based on data mining were applied. In the proposed approach, machine learning algorithms are used to distinguish between normal and attack traffic, and the most up-to-date data set are taken as reference for training the network. The simulation results have shown that the proposed model has the performance and reliability that can be used for WSNs.