by Ashis Sinha
In a major scientific breakthrough, researchers at Carnegie Mellon University (CMU) have developed a new artificial intelligence (AI) system that can track human movement in 3D — using only standard Wi-Fi routers. Remarkably, it works even when people are behind walls or in dark rooms, without the need for cameras.
This new system was developed by Jiaqi Geng, Dong Huang, and Fernando De la Torre, and is based on an advanced AI technique called DensePose, originally used in computer vision to map the human body in images. The CMU team adapted this method to work with Wi-Fi signals instead of cameras.
The study published with title “DensePose from WiFi”, shows how existing, everyday technology — like Wi-Fi — can be transformed into powerful tools for sensing and monitoring human activity. This opens up exciting possibilities in healthcare, smart home systems, and even entertainment. But it also highlights the need for responsible innovation.
How It Works
Wi-Fi signals naturally travel around our homes, bouncing off walls, furniture, and people. When someone moves, the way these signals bounce changes. The system captures these changes and uses a neural network — a type of AI model — to analyze the reflections.
Here’s the process:
- Three Wi-Fi routers are placed in a room.
- These routers send out signals that reflect off people’s bodies.
- The AI model reads the changes in signal phase and strength (known as Channel State Information, or CSI).
- Using this data, it recreates a 3D model of the person’s posture and movement in real time, even if the person is in another room or behind a wall.
This method is surprisingly accurate — it performs almost as well as traditional camera systems. And the best part? It works with inexpensive, off-the-shelf Wi-Fi hardware, like the TP-Link Archer A7 router, which costs around $50.
Potential Uses Across Different Fields
This technology opens the door to a wide range of real-world applications:
- Healthcare and Elderly Monitoring
-
- Track breathing, falls, and daily activity without cameras or wearables.
- Help monitor patients with mobility issues or chronic conditions.
- Useful in nursing homes or hospitals for passive, non-intrusive care.
- Smart Homes
-
- Adjust lights, heating, or appliances based on people’s movements.
- Enable privacy-preserving activity detection — no cameras needed.
- Security and Surveillance
-
- Detect intruders, loitering, or unusual behavior through walls or in darkness.
- Provide an alternative to visual surveillance, which raises privacy concerns.
- Gaming and Virtual Reality
-
- Allow for markerless motion capture — no need for suits or sensors.
- Enable more immersive experiences in AR/VR by tracking natural body movement.
Why It’s Important
Advantages over traditional methods:
-
- Privacy-friendly: Unlike cameras, Wi-Fi sensing doesn’t capture images or identifiable visuals.
- Low-cost and accessible: No need for expensive sensors like LiDAR or radar.
- Works in all environments: Functions in dark, crowded, or obstructed spaces.
Ethical and Privacy Concerns
Despite its many benefits, this technology raises serious questions:
-
- Surveillance risk: Someone could monitor people without their knowledge or consent.
- Legal gray areas: There are no clear laws governing how this type of Wi-Fi-based sensing can be used.
- Data security: Information gathered could be misused if not properly protected.
The researchers acknowledge these risks and stress the need for clear rules and safeguards. They suggest policies that include:
-
- Transparency: People should know if this technology is being used.
- Consent: Users should agree before being monitored.
- Ethical oversight: Independent bodies should oversee how it’s used.
- Privacy protections: Techniques like encryption or differential privacy should be explored.
The Big Picture
As this technology moves from the lab into homes and businesses, we must balance progress with privacy, ensuring that it helps society without being misused.
source: https://arxiv.org/pdf/2301.00250
