Artificial Neural Network Machine Learning Model for Malicious Action Detection and Prevention of MANET Assets

The main objective of this paper is to create a novel architecture of a machine learning model to identify and detect the malicious attacks occur in MANET statically and dynamically. Earlier research works have proposed various secured routing protocols for malicious activity detection whereas they are all detecting the attacker after the malicious action created in the network. However, the problem is, the MANET needs a novel method to identify and detect the malicious action to prevent the network, data, and other properties of the network. Hence this paper proposed an Artificial Neural Network model for learning, diagnosing, and detecting abnormal activities present in the MANET, and eliminate the corresponding actions and the node from the network function. Three different dataset is created in the network before and after deploying the nodes in the network, routing table, and intermediate node information input to ANN for analyzation. The simulation is created in NS-2 software and the results are verified based on the network performance with respect to remaining energy, throughput, packet delivery ratio, and packet loss. The performance of the ANN is evaluated by comparing its output with the other earlier methods and prove its efficiency.