The CSI ratio is a complex-valued signal derived by computing the element-wise quotient of CSI measurements from two adjacent antennas on the same WiFi receiver, effectively canceling out shared hardware impairments such as phase offsets and frequency-dependent noise that corrupt raw CSI readings. By eliminating these device-specific distortions, the CSI ratio yields a cleaner, more stable representation of the wireless channel, improving the reliability and generalizability of sensing tasks such as gesture recognition, localization, and human activity detection. It has been adopted as a base signal in deep learning benchmarks like SenseFi, where it serves as an alternative input modality to raw amplitude or phase CSI, with its primary variant being the pairwise antenna combination chosen from the available receive antenna array.
Source Papers
- Boosting WiFi Sensing Performance via CSI Ratio ↗ — Boosting WiFi Sensing Performance via CSI Ratio
- Deep Learning-Enhanced Human Sensing with Channel State Information: A Survey ↗ — Deep Learning-Enhanced Human Sensing with Channel State Info
- SenseFi: A library and benchmark on deep-learning-empowered WiFi human sensing ↗ — SenseFi: A library and benchmark on deep-learning-empowered