5-7-2023 (PITTSBURGH) Researchers at Carnegie Mellon University have successfully repurposed everyday WiFi routers into a groundbreaking tracking technology capable of detecting and monitoring the three-dimensional shape and movements of human bodies within a room. This innovative approach eliminates the need for cameras or expensive LiDAR hardware, relying solely on radio signals emitted by WiFi routers.
The team leveraged WiFi signals to create a substitute for RGB images in human sensing applications. Unlike traditional camera-based solutions, WiFi-based monitoring is minimally affected by factors such as illumination and occlusion. Additionally, this technology safeguards individuals’ privacy while remaining cost-effective, as most households in developed countries already possess WiFi routers. The researchers believe it can be scaled to monitor the well-being of elderly individuals or identify suspicious activities within homes.
In their study, which has yet to undergo formal peer review but is available on the preprint server ArXiv, the authors stated, “We believe that WiFi signals can serve as a ubiquitous substitute for RGB images for human sensing in certain instances. Illumination and occlusion have little effect on WiFi-based solutions used for interior monitoring. In addition, they protect individuals’ privacy and the required equipment can be bought at a reasonable price. In fact, most households in developed countries already have WiFi at home, and this technology may be scaled to monitor the well-being of elder people or just identify suspicious behaviors at home.”
To achieve their results, the researchers utilized DensePose, a system developed by Facebook’s AI lab and a London-based team. DensePose allows for the precise mapping of human body pixels, enabling the identification of key points and areas, including joints, arms, head, and torso. Combining this data with a deep neural network, the team mapped the phase and amplitude of WiFi signals transmitted and received by routers to specific coordinates on human bodies.
In a demonstration of their technology, the researchers employed three $30 WiFi routers and three aligned receivers that bounced WiFi signals off room walls. By isolating signals reflected off moving objects and disregarding static objects, the system reconstructed the pose of a person in a radar-like image. Remarkably, the method proved effective even when there was a wall obstructing the direct line between routers and subjects.
This novel approach potentially enables standard WiFi routers to see through various opaque obstacles, including drywall, wooden fences, and concrete walls.
While researchers have previously explored methods to “see” individuals through walls, such as using cell phone signals or WiFi, the Carnegie Mellon study offers significantly higher spatial resolution. The technology captures detailed poses of people in motion, providing valuable insights into their activities.
In a separate endeavor, Carnegie Mellon researchers previously developed a camera system capable of capturing sound vibrations with remarkable precision. This breakthrough allowed the system to reconstruct the music of individual instruments in a band or orchestra, without the need for microphones.
The researchers believe that WiFi signals can serve as a ubiquitous substitute for traditional RGB cameras, citing several advantages, such as the widespread presence of WiFi devices, their affordability, and the ability to overcome challenges faced by regular camera lenses, such as poor lighting and occlusion. They suggest that WiFi-enabled human detection can be used to identify and flag “suspicious behaviors” within households.
However, concerns surrounding personal privacy arise with the widespread adoption of WiFi-based human detection technology. As companies like Amazon explore the use of camera drones within homes, it is crucial to carefully consider the implications and potential breaches of privacy associated with this technology before it becomes commonplace in the market.
