South Korea has introduced a new innovative framework, MSF-Net, that improves the security of smart homes by using artificial intelligence and Internet of Things (IoT) technologies integrated into conventional WiFi infrastructure.

Researchers at Incheon National University have developed a system that can identify human activities in a room and automatically respond to potential threats or situations that require intervention. For example, MSF-Net can call emergency services if the owner falls or detect an attempted break-in. It can also adapt the functionality of a smart home, changing lighting or playing music according to the resident’s activities.
The system uses deep learning to analyze WiFi signals and recognizes a wide range of activities, including cooking or exercising. Scientist Gwang-gil Jeong emphasizes that
demonstrating high accuracy, including 91,82% on SignFi test datasets. This technology significantly improves the stability of WiFi systems by reducing the impact of interference on recognition accuracy.
Key components of MSF-Net include a two-stream framework for anomaly detection, transformers for extracting high-level features, and an attention mechanism that integrates data from different sources. An important advantage is that it can operate without additional sensors, which simplifies implementation into existing infrastructure. The researchers are also considering applications of this technology in the medical field, where it could help patients with disabilities or monitor the condition of patients in hospitals.
MSF-Net has attracted the attention of experts in the field of smart home security because it provides a high level of protection without the need to install cameras or motion sensors. It offers greater privacy to users and significantly reduces the cost of additional equipment. Experts expect that in the future this technology will become part of commercial solutions for smart homes, increasing their safety and comfort.