A Cooperative Attack Detection Framework for MANET-IoT Network Using Optimized Gradient Boosting Convolutional Neural Network

Abstract

The implementation of large-scale Internet of Things (IoT) devices results in smart cities. Using standard mobile ad-hoc networks and IoT, developers establish the communication model for a smart city. The rapid growth of IoT devices based on smart cities poses different Quality of Service (QoS) and security problems. This research presents a novel Modified Elephant Herd Optimization (MEHO) method and a Gradient Boosting Convolutional Neural Network (CNN) strategy to address these issues. The cooperative attacks with varied disruption probabilities are initially assessed at the edge nodes of the IoT network. The MEHO-based Gradient Boosting CNN (MEHO-CNN) approach effectively detects cooperative attacks, ensuring the identification of malicious entities. For traditional cloud access, both bandwidth utilization as well as expected latency are minimized in edge computing. By using the IoT network, the proposed MEHO-CNN model identifies and eliminates malicious nodes. To establish the claimed trustworthy background, the legitimate accusations are based on an examination of trust-based allegations. When compared to existing methodologies, the proposed approach lowers the impact of cooperative attacks, resulting in increased throughput, reduced attack detection rates, lower packet loss ratio, lower packet delivery ratio, and other benefits.