The Short-Time Fourier Transform (STFT) is a signal processing technique that divides a non-stationary CSI time series into short, overlapping temporal windows and applies the Fourier Transform to each segment, producing a time-frequency representation that captures how spectral content evolves over time. In WiFi-based human sensing, STFT is valuable because human activities such as walking, gestures, or breathing induce time-varying Doppler shifts and amplitude fluctuations in CSI signals that cannot be adequately characterized by either purely temporal or purely spectral analysis alone. A common variant used in this domain is the spectrogram, which visualizes the squared magnitude of the STFT output and serves as a two-dimensional feature map suitable for input into image-based classifiers such as convolutional neural networks, enabling robust discrimination across single-user and multi-user activity recognition scenarios.
Source Papers
- A Survey on Human Behavior Recognition Using Channel State Information ↗ — A Survey on Human Behavior Recognition Using Channel State I
- Channel State Information from Pure Communication to Sense and Track Human Motion: A Survey ↗ — Channel State Information from Pure Communication to Sense a
- Deep Learning-Enhanced Human Sensing with Channel State Information: A Survey ↗ — Deep Learning-Enhanced Human Sensing with Channel State Info
- WiMANS: A Benchmark Dataset for WiFi-based Multi-user Activity Sensing ↗ — WiMANS: A Benchmark Dataset for WiFi-based Multi-user Activi