Empowering Healthy Living & Frailty Detection with Privacy-Preserving Smart Activity Sensor​

An Initiative by AltumView x Keihanna Science City

image 4

What is the Proof of Concept (PoC)?

As part of the Keihanna EXPO Startup Challenge, AltumView proposed to extend the capabilities of AltumView’s Sentinare sensor. Instead of simply detecting incidents like falls, we will leverage AI and region-based movement analysis to:

  • Track behavioral patterns
  • Summarize health trends
  • Predict potential health risks

Core Shift: From incident detection to incident prevention

image

Goals of the Project

  • Use LLM and LLVM to interpret home movement data
  • Identify regions of interest (e.g., bathroom, bedroom) and flag anomalies
  • Track changes in daily routines and mobility to indicate declining health
  • Provide non-intrusive, AI-based summaries for caregivers and family
image 1

How Sentinare Compares

image 3

Current Progress

  • ✅ Participants successfully recruited at Club Keihanna
  • ✅ Sensors installed in real home environments
  • ✅ AI backend implemented with LLM and LLVM
  • ✅ Phase 1: LLVM is now actively verifying fall scenarios (fall detection accuracy improved)

What’s Next?

  • Real-time analysis through May–July
  • Integration of data into LLM for analysis and reporting
  • Showcase results at the Keihanna Festival – July 25, 2025
  • Prepare for large-scale rollout proposals in Fall 2025

Stay in Touch

Interested in the future of elder care, AI health trends, or smart home monitoring?