Systematic Literature Review of Intrusion Detection and Classification in Edge Computing: Types, Challenges, Solutions, Limitations and Research Directions

Abstract

This study presents a systematic literature review of intrusion detection and classification method for edge computing environment. Following PRISMA guided procedure, appropriate studies were identified from 2019 to 2026 through structured search method across relevant digital libraries, followed by thorough inclusion and exclusion screening. This review covered intrusion detection system (IDS) types and deployment structure. It also examined machine learning and deep learning method, feature engineering method, dataset, performance measures, and implementation. The review shows that host-based and anomaly-based intrusion detection system (IDS) lead in edge deployment. They track behaviors in detail and use few computing resources. Lightweight machine learning model like decision trees, random forests, and ensemble classifier are widely adopted, while Deep learning model often run into problems from limited resources. The review draws on this result to highlight key research gaps. It suggests paths ahead that focus on lightweight and federated detection system, standard test dataset, and low resource adaptive learning. It serves as a full guide for researchers and practitioners. Aim to build strong, scalable, and smart intrusion detection systems for safe edge computing environment.