Indiatimes iconIndiatimesSep 27, 2026 ~7 min source read

California teens build SafeStrides: a smartphone-and-wearable system that flags fall risk in older adults

Ruoqi Li, 16, and Jason Yang, 17, developed SafeStrides — a multimodal system using smartphone video plus wearable sensors to screen gait and mobility at home; their project won the 2026 Davidson Fellows engineering award and a shared $100,000 scholarship.

California teens, 16 and 17, build AI system to detect fall risks in older adults, win $100,000

Share this story

Send the public story page.

Useful takeaways from this story.

SafeStrides combines a smartphone camera with wearable sensors (IMUs and pressure insoles) to produce a multimodal fall-risk assessment that works at home.

The project progressed through three engineering generations: an Arduino prototype, a real-time IMU/ESP32 prototype, and a compact wearable plus a cross-platform Flutter app.

SafeStrides aims to detect changes months before a fall, giving older adults practical mobility and home-safety guidance and reducing reliance on infrequent clinic screenings.

# What SafeStrides is and why they built it

# How the system works SafeStrides pairs a smartphone app with wearable sensors to measure walking patterns and weight distribution. During a short guided test the app records video while the wearables stream sensor data. The key components are:

  • Inertial measurement units (IMUs) to capture motion data.
  • Pressure-sensing insoles to measure how weight is placed on each foot.
  • A smartphone camera that records gait and synchronises with the sensor streams.
  • Multimodal AI that analyses combined video and sensor inputs and returns a fall-risk evaluation plus mobility and home-safety suggestions.

The app synchronises high-frequency sensor streams with live camera footage so the multimodal model can examine how movement and foot pressure relate to balance and gait changes.

# Engineering progression Li and Yang iterated through three major hardware/software versions:

  1. Arduino-based prototype: functional but bulky, with data processed after tests rather than in real time.
  2. Real-time prototype: replaced the bulky components with IMUs, pressure insoles, an ESP32 microcontroller and Bluetooth modules to transmit sensor data live.
  3. Wearable design and app integration: refined the physical design to make the wearable compact and lightweight and built a cross-platform mobile application using Flutter that pairs with the hardware and synchronises data automatically.

# Intended use and practical benefits SafeStrides is designed for regular at-home screening so that changes in gait or balance can be detected earlier than typical clinic-based schedules. The developers point to several limitations of conventional assessments: travel to medical facilities, dependence on physician availability, and annual screening intervals that can miss rapid declines in mobility. SafeStrides offers:

  • Regular, shorter screenings that can be performed at home.
  • Immediate risk assessment and actionable guidance on mobility and home safety.

# What the award recognizes Fellows in engineering and awarded them a shared $100,000 scholarship for SafeStrides. The recognition highlights the project's combination of engineering development and a clear application to elder safety.

# What remains unreported in the story The published summary describes the system architecture, prototype generations, and the Davidson award but does not provide detailed validation results, deployment trials, regulatory status, or commercial availability. The article also does not include quantitative accuracy, sample sizes, or timeline for broader testing or rollout.

More context around this story.

Emergent Ventures winners, 59th cohort

Emergent Ventures winners, 59th cohort

Tym Syrytczyk, London, autonomous vehicles in the UK. Shane Regan, Long Island, 16, AI agents. Maximilian Kornstein, 15, Atlanta area, agents and general career support. Irene Chen and Jessica Dai, UC Berkeley, data on peptides use. Evan Warfel, Bay Area, updated meta-analyses through AI. Daniel Dominguez Gomez, biomed

Loading more related stories...

Keep reading in the app

Open the app view to save this story, compare related coverage, and continue from the same source.

Open in app