CSI-based activity recognition refers to the use of Channel State Information (CSI) extracted from Wi-Fi signals to identify and classify human physical activities, such as walking, sitting, falling, or gesturing, by analyzing the fine-grained amplitude and phase perturbations that body movements induce on multipath wireless propagation. It matters because it enables passive, non-intrusive, and privacy-preserving sensing without requiring subjects to carry dedicated devices, making it attractive for applications in healthcare monitoring, smart homes, and security. Key variants include gesture recognition, gait identification, fall detection, and more complex multi-person or continuous activity recognition, with ongoing research addressing generalizability challenges such as cross-environment, cross-subject, and cross-device deployment to ensure robust real-world performance.
Source Papers
- A CSI Dataset for Wireless Human Sensing on 80 MHz Wi-Fi Channels ↗ — A CSI Dataset for Wireless Human Sensing on 80 MHz Wi-Fi Cha
- A Survey on Wi-Fi Sensing Generalizability: Taxonomy, Techniques, Datasets, and Future Research Prospects ↗ — A Survey on Wi-Fi Sensing Generalizability: Taxonomy, Techni