SMVT-WFF:A lightweight IoT intrusion detection model based on improved MobileViT
Shunhong Long, Yong Wang · 2025
In the realm of cybersecurity, the Internet of Things (IoT) has become increasingly significant due to the widespread integration of network-connected devices in daily life. Intrusion Detection Systems (IDS) for IoT are crucial for safeguarding these interconnected devices against malicious attacks and unauthorized intrusions. Given that IoT environments often face resource constraints, such as limited processing power and minimal memory, the development of lightweight intrusion detection models is particularly important. These models must effectively identify and mitigate threats without significantly increasing the computational resource burden or depleting the limited energy reserves of the devices. In this study, we propose a novel IoT intrusion detection model, SMVT-WFF, which effectively detects network attacks in resource-constrained IoT environments. This model streamlines the structure of MobileViT and introduces a Weighted Feature Fusion (WFF) module to capture multi-scale spatial information, significantly enhancing feature extraction capabilities. Experimental results demonstrate that the model achieves an accuracy of 99.87% on the BoT-IoT dataset. Compared to traditional lightweight models, it exhibits notable performance advantages, as confirmed by various evaluation metrics.