A Lightweight Intrusion Detection System for Unmanned Aerial Vehicles Using CNN

View at publisher Summarise with VIRAT
Authors: Hyungchul Im, Sukju Kim, Seongsoo Lee
Cited by: 0
References: 17

Summary

No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.

Summarise with VIRAT Free to try with your Science Catalogs account.

Abstract

Unmanned aerial vehicles (UAVs), commonly known as drones, are used for tasks such as delivery, agriculture, and observation. However, with the increasing adoption of UAVs, the risk of cyberattacks exploiting vulnerabilities in the controller area network (CAN) protocol also rises. CAN is an in-vehicle network that transmits data between electronic control units (ECUs). Recently, various intrusion detection systems (IDSs) for CAN have been proposed; however, these vehicle-oriented IDSs are not directly applicable to uncomplicated application-level vehicular computing and networking (UAVCAN), the higher-layer protocol used in UAVs. Therefore, we propose a convolutional neural network (CNN)-based IDS designed specifically for UAVCAN and for deployment on resource-constrained UAV platforms. The proposed model captures the message structure and characteristics of UAVCAN by using data sequences as input features, whereas CAN-specific IDSs typically use ID sequences. We evaluate our model on data collected from real drones using UAVCAN. The experimental results show that our IDS achieves superior detection performance compared with CAN-specific IDSs and highlight the need for a UAVCAN-specific solution.

Fields of Study & Subjects

Network Security and Intrusion Detection Adversarial Robustness in Machine Learning Smart Grid Security and Resilience

Publication details

Journal / source
International Conference on Consumer Electronics
Pages
1–5
Publication date
3 February 2026
Language
EN