Kevin Tang, a 13-year-old from California, has transformed personal tragedy into technological triumph with FallGuard—an AI-powered fall detection system that just won the top prize at the 2025 3M Young Scientist Challenge. Unlike commercial wearables, his privacy-focused, camera-based solution offers real-time alerts to multiple caregivers without subscription fees, and he’s already used his $25,000 prize to launch it as a free public app.
The story of FallGuard begins not in a lab, but in a kitchen. A few years ago, Kevin Tang’s grandmother fell, and the family didn’t notice immediately. By the time they found her and called for help, she was left with permanent brain damage. Later, Tang learned a friend’s grandparent had also fallen, undiscovered for a full day because they lived in another state.
These incidents illuminated a brutal reality: falls are a leading cause of injury and death among older adults, with the CDC reporting millions of such incidents annually. For Tang, this wasn’t just a statistic; it was a call to action. He embarked on a mission to build a solution that could prevent such outcomes for other families.
How FallGuard’s AI Sees What Humans Miss
Starting in the summer of 2024, Tang developed FallGuard into a sophisticated system that differs fundamentally from existing market solutions. Instead of a wearable device that requires charging and remembering to put it on, FallGuard uses a wall-mounted camera connected to a computer running his custom AI model.
The system leverages Google’s MediaPipe, an open-source framework for building multimodal applied ML pipelines, to map a person’s body with key points. Tang’s own two-stage algorithm then analyzes posture and movement over time. It uses bounding boxes—a common computer vision technique—to track how a person’s body proportions change from a standing to a lying position.
What makes it intelligent is its velocity analysis: if the AI detects a lay-down event, it checks the previous second to determine if the person’s movement suddenly dropped, distinguishing an accidental fall from someone intentionally lying down.
When a fall is detected, the system immediately sends an alert to family members’ phones through the FallGuard mobile app. Crucially, no video is recorded or uploaded, addressing significant privacy concerns that often plague camera-based monitoring systems. The system doesn’t rely on cellular carriers and generates no messaging fees, making it accessible without ongoing costs.
From Science Fair to Real-World Impact
At the 2025 3M Young Scientist Challenge, Tang was paired with Mark Gilbertson, a robotics and AI specialist at 3M, who provided mentorship on practical implementation questions—like how to mount the device and what materials to use. Gilbertson noted that Tang’s project stood out specifically because of its emotional connection to real-world problems.
Tang’s $25,000 prize wasn’t just recognition—it became development capital. He immediately reinvested part of it into improving FallGuard, including purchasing a MacBook to code the FallGuard computer app that allows anyone to convert their existing computer into a fall detection device.
The response has been overwhelming. Approximately 500 families have expressed interest in FallGuard. One particularly poignant inquiry came from a deaf man struggling to care for his wife. As Gilberston relayed, the man explained that “This invention will just really change our lives and quality of living,” highlighting how Tang’s technology addresses needs that conventional solutions miss.
The Technical Frontier and What’s Next
Like any version 1.0 technology, FallGuard has limitations. The current system requires falls to occur within the camera’s field of view, and each computer can only support one camera at a time. Tang is already working on expanding the system to support multiple cameras throughout a home without requiring multiple computers.
He’s also refining the AI’s reliability to reduce false positives and ensure it only alerts when actual falls occur. This work places him at the forefront of a growing movement to use affordable computer vision for in-home safety monitoring, a field typically dominated by expensive subscription services and less reliable wearable buttons.
When asked what he’s most proud of, Tang doesn’t mention the prize money or media attention. He points to the evolution of the project itself—from a tripod and camera to a mounted device, and finally to an app that anyone can download. “I just kept working until I had a final product,” he said, embodying the iterative spirit of true innovation.
For the latest developments in how AI is transforming healthcare and home safety, continue reading on onlytrustedinfo.com, where we provide the fastest, most authoritative analysis of breaking technology news.
