A simple radio receiver can detect human motion and estimate walking speed by analyzing the interference patterns of wideband signals, such as those used for digital TV. By monitoring how a signal's phase is disrupted by physical objects, it is possible to identify movement even through walls, a phenomenon the author of a recent project describes as 'happy paranoia.'
The Mechanism of Selective Fading
The primary source describes a phenomenon called selective fading, where a direct signal is canceled out by a strong reflection. Because the phase of these reflections depends on frequency, the effect is not uniform across the entire signal.
This image provides a visual baseline of a stable radio signal, which is the starting point for observing the selective fading patterns caused by human movement.
This state is extremely sensitive to any nearby movement. When a person walks through the signal path, they cause a slight disturbance that manifests as distinct stripes in a spectrogram, or 'waterfall' plot.
The gaps between these stripes occur specifically when the reflection perfectly cancels out the residual signal, which happens each time the path length changes by exactly one wavelength.
“The gaps between the stripes happen when my reflection perfectly cancels out the residual signal, which happens each time the path length changes by one wavelength.”
— Your TV can see through walls:
Calculating Movement from Radio Waves
The author demonstrated this by recording an ATSC digital TV signal using an Ettus USRP B200. By observing the signal while walking in a neighboring room, they were able to visualize the movement as a series of stripes.
To calculate speed, the author measured the gap between stripes (approximately 70 pixels) and compared it to the waterfall's update rate (150 Hz). This allowed them to determine a walking speed of 1.3 m/s, or roughly 4.6 km/h.
The author notes that this method applies to any wideband radio signal, including WiFi or LTE. However, these signals often produce a 'messier' waterfall, requiring more effort to extract the same data.
Community Discussion and Broader Context
In the Hacker News discussion, commenters noted that this is not a new concept; some pointed out that analog terrestrial antennas have historically been affected by physical proximity, a fact often used in physical comedy.
One commenter, writing as numlock86, suggested that similar principles apply to fiber-optic cables, where sound waves can 'bend' the cable enough to be measured and transformed back into audio signals. Others mentioned that RF signals have been used in studies to estimate traffic flow using 4G data.
The discussion also highlighted the privacy implications, with one commenter, writing as ThePumpkinGod, noting that a bad actor could use these methods to determine if someone is home. Another commenter, writing as darkwater, questioned why such 'magic' presence detection isn't already integrated into smart home systems like Home Assistant.