Intrusion detection in MANET using Self Organizing Map (SOM)

Mobile Ad-hoc networks (MANET) are formed with dynamism and upheld by individual hosts in a network. In these type of networks all communication occurs through a wireless medium and the nature of the network is decentralized and dynamic. Hence it probes for a number of security problems and in order provide security against malicious attacks, Intrusion Detection System (IDS) is commonly used as a second route of protection in MANET. Intrusion detection models are used to detect the attacks based on the patterns and alerts in case of intruders are being met with the system. In this paper, we propose and implement intrusion-detection system grounded on artificial neural network model such as Self-Organizing Map (SOM) based competitive network, which in turn plays a vital role in detection of malicious nodes based on input data patterns. The proposed model deals with different types of attacks and their detection approach based on SOM model. The approach aids at increasing Detection rate as well as reducing the False alarm rate, which in turn helps to detect those attacks before it makes larger damage to the network and prevent them with supportive techniques and increase the network performance. The experimental results of proposed model is evaluated under different parameters.