Open Access

ARTICLE

Droid-IoT: Detect Android IoT Malicious Applications Using ML and Blockchain

Hani Mohammed Alshahrani^*

College of Computer Science and Information Systems, Najran University, Najran, 61441, Saudi Arabia

* Corresponding Author: Hani Mohammed Alshahrani. Email:

(This article belongs to this Special Issue: Emerging Trends in Artificial Intelligence and Machine Learning)

Computers, Materials & Continua 2022, 70(1), 739-766. https://doi.org/10.32604/cmc.2022.019623

Received 20 April 2021; Accepted 24 May 2021; Issue published 07 September 2021

Abstract

One of the most rapidly growing areas in the last few years is the Internet of Things (IoT), which has been used in widespread fields such as healthcare, smart homes, and industries. Android is one of the most popular operating systems (OS) used by IoT devices for communication and data exchange. Android OS captured more than 70 percent of the market share in 2021. Because of the popularity of the Android OS, it has been targeted by cybercriminals who have introduced a number of issues, such as stealing private information. As reported by one of the recent studies Android malware are developed almost every 10 s. Therefore, due to this huge exploitation an accurate and secure detection system is needed to secure the communication and data exchange in Android IoT devices. This paper introduces Droid-IoT, a collaborative framework to detect Android IoT malicious applications by using the blockchain technology. Droid-IoT consists of four main engines: (i) collaborative reporting engine, (ii) static analysis engine, (iii) detection engine, and (iv) blockchain engine. Each engine contributes to the detection and minimization of the risk of malicious applications and the reporting of any malicious activities. All features are extracted automatically from the inspected applications to be classified by the machine learning model and store the results into the blockchain. The performance of Droid-IoT was evaluated by analyzing more than 6000 Android applications and comparing the detection rate of Droid-IoT with the state-of-the-art tools. Droid-IoT achieved a detection rate of 97.74% with a low false positive rate by using an extreme gradient boosting (XGBoost) classifier.

---

Keywords

---