Jamming Detection in IoT Wireless Networks: An Edge-AI Based Approach
Author | Hussain, Ahmed |
Author | Abughanam, Nada |
Author | Qadir, Junaid |
Author | Mohamed, Amr |
Available date | 2023-05-22T10:48:29Z |
Publication Date | 2022-11-07 |
Publication Name | ACM International Conference Proceeding Series |
Identifier | http://dx.doi.org/10.1145/3567445.3567456 |
Citation | Hussain, A., Abughanam, N., Qadir, J., & Mohamed, A. (2022, November). Jamming Detection in IoT Wireless Networks: An Edge-AI Based Approach. In Proceedings of the 12th International Conference on the Internet of Things (pp. 57-64). |
ISBN | 978-145039665-3 |
Abstract | Wireless enabling technologies in critical infrastructures are increasing the efficiency of communications. In the era of 5G and beyond, more technologies will be allowed to connect to mobile networks, enabling the Internet of Things (IoT) on a massive scale. Most of these technologies are vulnerable to physical-layer security attacks, namely jamming. Jamming attacks are among the most effective techniques to attack and compromise the availability of these wireless technologies. Jamming is an interfering signal that limits the intended receiver from correctly receiving the messages. Once the adversary deploys a jammer in a wireless network, jammer detection becomes difficult, if not impossible, due to the inaccessibility of the affected devices in the network. This paper extends the state-of-the-art jamming detection and classification methods by proposing an effective IoT Tiny Machine Learning (TinyML)-based approach, where a trained deep learning model is deployed on an IoT edge device, namely a Raspberry Pi. The model is built using TensorFlow and deployed on the IoT device using TensorFlow lite. The trained model encompasses two commonly known jamming types: constant and periodic, in addition to the normal channel state. The Raspberry Pi is connected to a Software Defined Radio (SDR) that continuously senses the WiFi channel and acquires Received Signal Strength (RSS) readings which the TinyML model evaluates to detect the presence of jamming and its type. We release both the procedure and collected dataset for the different types of jamming as open source. Finally, we conducted an extensive testing campaign to test, evaluate, and illustrate the effectiveness of the proposed TinyML-based detection on the edge scheme. |
Sponsor | This work was made possible by Qatar National Research Fund (a member of Qatar Foundation), NPRP grant # NPRP12S-0119-190006. |
Language | en |
Publisher | ACM Digital Library |
Subject | Deep Learning Edge AI Internet of Things Jamming Tensorflow TinyML Wireless Communication |
Type | Conference Paper |
Pagination | 57-64 |
Files in this item
This item appears in the following Collection(s)
-
Computer Science & Engineering [2288 items ]