A HYBRID FRAMEWORK FOR DETECTION AND MITIGATION OF DDOS ATTACKS IN IOT ENVIRONMENT

Abstract

Internet of Things (IoT) devices are becoming increasingly valuable, but they moreover come with more security dangers and vulnerabilities because of their constrained assets. Distributed Denial of Service(DDoS) could be a noteworthy peril to IoT gadgets. This paper presents a novel learning design for distinguishing and mitigating DDoS attacks in IoT systems. It uses a Robust covariance-based Principal Component Analy- sis (RCPCA) technique for pre-processing, which successfully identifies and removes outliers from the complex IoT network traffic data. The Lyrebird Armadillo Fusion Algorithm (LAFA), which combines the Lyrebird Optimization Algorithm (LOA) with the Giant Armadillo Optimization (GAO), is used by the framework to select optimal features. The Proposed Framework presents the Capsule Gated Perception (CGP) model, a deep learning model that combines an optimized Multi-Layer Perceptron (MLP), Gated Repetitive Units (GRU), and Capsule Systems for the purpose of assault discovery. In conclusion, the proposed framework employments Profound Q-Networks (DQN) to mitigate assaults by reacting to DDoS assaults in real- time and reinforcing the security and flexibility of IoT systems.