CSI-based WiFi sensing is a technique that extracts Channel State Information (CSI) from WiFi signals to infer physical phenomena such as human activity, gesture recognition, and localization, by analyzing how the wireless channel response is perturbed by changes in the environment. It matters because it enables contactless, privacy-preserving sensing using existing WiFi infrastructure, making it a scalable alternative to camera-based or wearable sensing systems. Key variants differ in the machine learning paradigms applied to CSI data, ranging from supervised deep learning architectures such as CNNs, ResNets, and RNNs to self-supervised approaches like Contrastive Predictive Coding, as well as in the specific sensing tasks targeted, including human activity recognition, indoor localization, and gesture detection.
Source Papers
- Context-Aware Predictive Coding: A Representation Learning Framework for WiFi Sensing ↗ — Context-Aware Predictive Coding: A Representation Learning F
- CrossSense: Towards Cross-Site and Large-Scale WiFi Sensing ↗ — CrossSense: Towards Cross-Site and Large-Scale WiFi Sensing
- SenseFi: A library and benchmark on deep-learning-empowered WiFi human sensing ↗ — SenseFi: A library and benchmark on deep-learning-empowered
- WiFi Sensing with Channel State Information ↗ — WiFi Sensing with Channel State Information
- WiGNN: WiFi-Based Cross-Domain Gesture Recognition Inspired by Dynamic Topology Structure ↗ — WiGNN: WiFi-Based Cross-Domain Gesture Recognition Inspired