A Review on IoT Cyber-Attacks Detection Challenges‎ and Solutions

Abstract

The Internet of Things (IoT) has recast the way we interact with technology and the world around us. The IoT is almost everywhere in our daily lives, and this huge growth in using the IoT increases the need for implementing a high level of security framework for this technology. This study presents an analysis of some of the recent IoT cyber-attack detection systems to provide an evaluation of these systems and useful future research directions in this field. This analysis is introduced along with an overview of the IoT's security challenges and solutions and the types of security attacks in the IoT environment. Although the most recent approaches to cyber-attack detection have high percentages of attack detection accuracy, these classical DL models that were learned with local datasets need more enhancement to maintain privacy and data storage when involved in networks of cooperative nodes in IoTs. Furthermore, current techniques still lack the ability to provide a generalization for new attacks, cover the binary and multiclass classifications of cyber-attacks, and develop feature selection algorithms to reduce dataset dimension.