Deep Learning based IDS for Secured Routing in Wireless Sensor Networks using Fuzzy Genetic Approach

Abstract

Wireless Sensor Networks (WSNs) are consisting of a number of sensing nodes that are used to collect data from various environmental conditions and are used frequently when it comes to monitoring and regulating of the network environment. In this scenario, the presence of malicious nodes in the network will lead to challenges in the provision of communication security since it is considered as a significant issue in the successful delivery of the acquired data. Therefore, it is necessary to safeguard the network communication from the security threats by identifying the dangerous behaviours in each of the sensor nodes and then by isolating the malicious nodes. This can be accomplished by the deployment of Intrusion Detection Systems (IDSs) at the sensor nodes. However, the existing security systems using IDSs are unable to attain sufficient detection accuracy even though the security models are developed using multiple strategies. To handle this issue, we have proposed a novel Intrusion Detection System (IDS) in this work by using a Convolution Neural Network based Deep Learning (DL) approach along with a fuzzy inference model. In this work, the proposed IDS keeps track of the network and system activities by using a Fuzzy Convolutional Neural Network (F-CNN) which is proposed in this paper to look for signs of malicious activities. In addition to this IDS, we propose a new A-Star search based Neuro-Genetic Secure Routing Algorithm (ASNGSRA) in order to increase the security of communication in WSNs. The proposed ASNGSRA framework also predicts the best optimal route in a secured way from the source to destination. The major advantages of the IDS based secured routing model with neuro-genetic route discovery process proposed in this paper include the increase in security, packet delivery ratio and network lifespan as well as the reduction in delay and energy consumption.

Blockchain-Based Secure Localization against Malicious Nodes in IoT-Based Wireless Sensor Networks Using Federated Learning

“…Data Preprocessing and Feature Selection. The traffic data can be collected from a local network or the benchmark dataset using the Data collection module [40]. The preprocessing module receives these data and filters them before they are used.…”