Abstract: An ad hoc mobile network is a set of autonomous nodes; these nodes can send and receive data independently. Security is a major concern for MANET because ad hoc networks are based on trust; each node of a network depends on its neighboring node, each node of a network works well as a router. Now, if a malicious node in that system is a great challenge for researchers. In this paper, we perform a detailed analysis of various types of attacks in mobile ad hoc networks (such as denial of service attacks, investigations, user attacks against root, vampire attacks, etc.) To protect the network against such vulnerabilities, a system capable of mitigating these attacks on the network is needed. Therefore, we have performed a detailed search on various types of intrusion detection systems. After studying the IDS, we conclude that all the above approaches have their advantages and disadvantages, but one thing that is common to all is that the detection rate of hybrid attacks is low, Therefore, we have proposed a new technique in which we use a support vector machine, as well as a dendritic cell algorithm that classifies abnormal and normal data traffic according to its acquired rules, as well as the predefined rules taught by this system. Likewise it is capable of difference between normal data and abnormal data. The proposed IDS approach is equipped with a learning algorithm used to form the support vector machine. Wireless network that achieves high accuracy to detect hybrid attacks, as well as normal and heterogeneous behaviors. The DCA and SVM classifiers will reach a detection rate of 100% (fixed duration).
Keywords: MANET, R2L, U2R, IDS, SVM, MAC
|
DOI:
10.17148/IJARCCE.2019.81002
