Research Article
Towards an Autonomous Host-Based Intrusion Detection System for Android Mobile Devices
@INPROCEEDINGS{10.1007/978-3-030-05195-2_14, author={Jos\^{e} Ribeiro and Georgios Mantas and Firooz Saghezchi and Jonathan Rodriguez and Simon Shepherd and Raed Abd-Alhameed}, title={Towards an Autonomous Host-Based Intrusion Detection System for Android Mobile Devices}, proceedings={Broadband Communications, Networks, and Systems. 9th International EAI Conference, Broadnets 2018, Faro, Portugal, September 19--20, 2018, Proceedings}, proceedings_a={BROADNETS}, year={2019}, month={1}, keywords={Mobile Intrusion Detection System Android Security 5G communications Machine Learning Malware detection Host-based IDS}, doi={10.1007/978-3-030-05195-2_14} }- José Ribeiro
Georgios Mantas
Firooz Saghezchi
Jonathan Rodriguez
Simon Shepherd
Raed Abd-Alhameed
Year: 2019
Towards an Autonomous Host-Based Intrusion Detection System for Android Mobile Devices
BROADNETS
Springer
DOI: 10.1007/978-3-030-05195-2_14
Abstract
In the 5G era, mobile devices are expected to play a pivotal role in our daily life. They will provide a wide range of appealing features to enable users to access a rich set of high quality personalized services. However, at the same time, mobile devices (e.g., smartphones) will be one of the most attractive targets for future attackers in the upcoming 5G communications systems. Therefore, security mechanisms such as mobile Intrusion Detection Systems (IDSs) are essential to protect mobile devices from a plethora of known and unknown security breaches and to ensure user privacy. However, despite the fact that a lot of research effort has been placed on IDSs for mobile devices during the last decade, autonomous host-based IDS solutions for 5G mobile devices are still required to protect them in a more efficient and effective manner. Towards this direction, we propose an autonomous host-based IDS for Android mobile devices applying Machine Learning (ML) methods to inspect different features representing how the device’s resources (e.g., CPU, memory, etc.) are being used. The simulation results demonstrate a promising detection accuracy of above 85%, reaching up to 99.99%.