Article Details

Abstract

To address the severe security threats faced by IoT edge devices in large-scale deployment, the poor adaptability of traditional protection methods, and the shortcomings of single detection mechanisms in identifying unknown threats and controlling the false alarm rate, this study investigates the IoT edge security protection system and proposes a protection scheme based on dual-engine collaborative detection and behavior modeling. The scheme integrates a static rule detection engine and a dynamic behavior analysis engine: the former rapidly identifies known attacks through rule base matching, while the latter builds a baseline of normal device behavior through machine learning to capture abnormal activities. The two engines exchange information through a designed collaboration mechanism and fuse detection results based on confidence. Meanwhile, strategies such as model lightweighting and computational optimization are adopted to adapt to the resource constraints of edge devices. Verified on an edge environment simulation platform built with a Raspberry Pi cluster and Mininet as well as on standard datasets, the proposed system achieves higher accuracy in detecting both known and unknown attacks with lower resource consumption, effectively improving the security protection capability of the IoT edge environment.

Keywords

dual-engine collaboration; behavioral modeling; IoT edge security; anomaly detection; rule detection; resource optimization

References

1. [1] Aditi R，Kokila J，Ramasubramanian N，et al. Device-Specific Security Challenges and Solution in IoT Edge Computing: A Review [J].The Journal of Supercomputing，2023，79（18）：20790-20825.
2. [2] Li Jin, Gao Hongliang, Liu Kemeng, etc Intrusion Security Protection of Power System Based on Edge Node Technology [J]. Modern Electric Power, 2024, 41 (6): 1009-1016
3. [3] Gu Tiansheng, Zeng Fukang, Shao Sisi, etc Intelligent IoT terminal security protection method based on mimetic defense and Paillier encryption [J]. Journal of Network and Information Security, 2024, 10 (6): 164-176
4. [4] Zhou Pinxi, Shen Yue, Li Wei Overview of IoT Intrusion Detection Systems Based on Deep Learning [J]. Network Security and Data Governance, 2025, 44 (6): 1-10
5. [5] Biswas A，Wang H C. Autonomous Vehicles Enabled by the Integration of IoT, Edge Intelligence, 5G, and Blockchain [J].Sensors，2023，23（4）：1963.
6. [6] Liu Huifei, Zhu Yulei, Liu Haotian, etc Overview of Industrial Internet of Things Security Protection Based on Artificial Intelligence [J]. Internet of Things Technology, 2025, 15 (10): 118-125
7. [7] Nie Zhen, Guo Zichen, Gao Ming, etc Mobile Communication Network Slice Security Deployment System Based on Machine Learning [J]. Electronic Design Engineering, 2023, 31 (17): 173-177
8. [8] Wang Kebing, Yang Guowei, Lian Changjin Research on the Security Mechanism and Strategy of 5G Network in Coal Mines [J]. China Mining, 2024, 33 (12): 173-179
9. [9] Ren Xiaoxu, Qiu Chao, Deng Hui, etc Edge Intelligence Fusion Blockchain: Research Status, Applications, and Challenges [J]. Information and Control, 2024, 53 (1): 1-16
10. [10] Feng Yixiong, Yang Chen, Hu Bingtao, etc Design and application of cloud based PLC system architecture based on 5G multi access edge computing [J]. Journal of Computer Aided Design and Graphics, 2024,36 (1): 33-46

Funding

Youth Innovative Talents Project for Ordinary Universities in Guangdong Province in 2025 (2025KQNCX230); University level research projects of Guangzhou City Vocational College (KYCT2025047, KYCT2025038, KYFB2025032)

About the authors

Huang Minfeng (2001-), male, Han ethnicity, from Shantou, Guangdong, teaching assistant, master's student, research interests include IoT application technologies and applied AI technologies; Yu Junyue (2006-), male, Han ethnicity, from Huizhou, Guangdong, research interests include network security and AI model security and its applications.