Artificial Ecosystem Optimizer with Convolutional Recurrent Neural Network for Intrusion Detection System in Wireless Sensor Networks

Abstract

Wireless Sensor Network (WSN) encompasses Sensor Nodes (SNs) used to gather data regarding surroundings. Some features that make WSNs vulnerable to security attacks are open wireless medium, distributed nature, and multi-hop data forwarding. Several security-based solutions for WSNs were devised, like secure routing or security mechanisms, authentication, and key exchange for particular attacks. Such security mechanisms can ensure security at a certain level but cannot eradicate various security attacks. Intrusion Detection System (IDS) was crucial in preventing and detecting security attacks. This study develops an Artificial Ecosystem Optimizer with Convolutional Recurrent Neural Network for IDS (AEOCRNN-IDS) in WSN. The main aim of the AEOCRNN-IDS model lies in recognising and classifying intrusions present in the network. Data preprocessing is done in the suggested AEOCRNN-IDS model to make the data compatible for further processing. The AEOCRNN-IDS approach uses the CRNN technique in this study for intrusion detection and classification. Since hit-and-miss tuning selection is a dreary process, the AEOCRNN-IDS approach implements the AEO method for optimal tuning. The experimental validation of the AEOCRNN-IDS system was tested employing an intrusion database from the Kaggle repository. The extensive experimental analysis stated the authority of the AEOCRNN-IDS approach on IDS in the WSN.