Recent Reinforcement Learning and Blockchain Based Security Solutions for Internet of Things: Survey

Rim Gasmi, Sarra Hammoudi, Manal Lamri, Saad Harous

Research output: Contribution to journalArticlepeer-review

Abstract

Users’ security is one of the most important issues in Internet of Things (IoT) due to the high number of IoT devices involved in different applications. Security threats are evolving at a rapid pace that make the current security and privacy measures unsuitable. Therefore, several researchers have been attracted by this domain with the aim of proposing either new or improved solutions to address the problem of security in IoT. Blockchain technology is a relatively new invention in modern IoT applications to solve the security issue. It is based on the use of a public immutable ledger called a blockchain. After conducting a verification process, several parts on a network encode transactions into this ledger. Moreover, Machine learning (ML) algorithms have been used as emerging solutions to improve IoT security. Reinforcement learning (RL) is the most popular machine learning technique proposed to secure IoT systems. Unlike other ML methods, RL can observe, learn and interact with the environment even if it has minimum information about the considered parameters. Various researches have been proposed to treat security problem in IoT based on either RL technique or Blockchain technology or a combination of both techniques. Therefore, we believe there is a need for a comprehensive survey on works proposed in recent years that address security issues using these techniques. In this paper, we provide a summary of research efforts made in the past few years, from 2018 to 2021, addressing security issues using RL and blockchain techniques in the IoT domain.

Original languageEnglish
Pages (from-to)1307-1345
Number of pages39
JournalWireless Personal Communications
Volume132
Issue number2
DOIs
Publication statusPublished - Sept 2023
Externally publishedYes

Keywords

  • Blockchain
  • Internet of Things
  • Machine learning
  • Reinforcement learning
  • Security

ASJC Scopus subject areas

  • Computer Science Applications
  • Electrical and Electronic Engineering

Fingerprint

Dive into the research topics of 'Recent Reinforcement Learning and Blockchain Based Security Solutions for Internet of Things: Survey'. Together they form a unique fingerprint.

Cite this