AI Fall Detection: Radar and Wi-Fi in Eldercare

Eldercare facilities face a constant challenge in preventing severe injuries from bathroom falls while respecting resident privacy. Because traditional optical cameras are intrusive and inappropriate for sensitive areas, facilities are adopting ambient sensing technologies. This article explores how nursing homes and assisted living centers combine millimeter-wave (mmWave) radar and Wi-Fi Channel State Information (CSI) with artificial intelligence to accurately detect falls in real time without recording a single visual image.

The Need for Camera-Free Monitoring

Bathrooms are the most common location for severe falls among the elderly due to slick surfaces and awkward transitions. However, installing video cameras in bathrooms violates resident dignity and often breaches healthcare privacy regulations. Wearable panic buttons also fall short, as residents frequently forget to wear them, take them off to shower, or lose consciousness during a fall. Ambient, non-visual sensors solve this by monitoring the physical space rather than relying on the resident's compliance or capturing identifiable imagery.

Millimeter-Wave (mmWave) Radar Technology

Millimeter-wave radar operates by emitting high-frequency electromagnetic waves into the bathroom. When these waves hit an object or a person, they bounce back to the sensor.

Instead of creating a photograph, the radar measures the time it takes for the waves to return and the frequency shift of those waves. This data generates a "3D point cloud"—a cluster of digital dots that represents the mass, height, and movement speed of a person in the room. Because the radar easily penetrates steam, water droplets, and darkness, it remains reliable even when a resident is taking a hot shower in a dimly lit room. If the point cloud suddenly drops from a standing height to the floor level at a rapid velocity, the system flags a potential fall.

Wi-Fi Channel State Information (CSI)

Wi-Fi CSI utilizes the standard wireless signals that already exist within an eldercare facility. As Wi-Fi waves travel between a transmitter and a receiver, they bounce off walls, fixtures, and people.

CSI measures how these signals are altered—specifically looking at changes in amplitude and phase—as they pass through the room. When a resident moves, their body disrupts the Wi-Fi waves in a unique way. While less spatially precise than mmWave radar, Wi-Fi CSI covers the entire room without blind spots. A sudden, violent disruption in the Wi-Fi signal pattern, followed by a complete lack of signal variation, indicates that a person has fallen and is lying motionless.

The Role of Artificial Intelligence

Raw data from radar point clouds and Wi-Fi disruptions is useless without interpretation. This is where artificial intelligence, specifically deep learning and machine learning algorithms, steps in.

AI models are trained on thousands of hours of simulated and real-world movement data. The AI learns to differentiate between a genuine fall and benign activities, such as a resident bending down to pick up a dropped towel, sitting quickly on a toilet, or a caregiver entering the room.

When a fall occurs, the AI processes the event in milliseconds:

  1. Detection: The AI identifies the rapid downward acceleration via radar and the sudden signal scatter via Wi-Fi.
  2. Verification: The system monitors the resident for a set period (usually 5 to 15 seconds) to see if they get back up.
  3. Vital Sign Check: Advanced mmWave radar is sensitive enough to detect the micro-movements of a person's chest, allowing the AI to monitor breathing and heart rate while the resident is on the floor.
  4. Alerting: If the resident remains on the floor, the AI instantly sends an alert to the nursing station or caregiver mobile devices, detailing the exact room and the severity of the fall.

Combining Forces for Maximum Safety

By fusing mmWave radar and Wi-Fi CSI, eldercare facilities eliminate the weaknesses of using either technology alone. Radar provides high-precision tracking and vital sign monitoring, while Wi-Fi CSI ensures whole-room coverage, even behind shower curtains or partitions. Most importantly, because the data consists entirely of radio wave reflections and mathematical coordinates, it is physically impossible to reconstruct a visual face or body from the data. Residents receive continuous, life-saving protection with 100% of their privacy intact.