Journal

Ambient Assisted Living: Sensor Fusion for Elderly Fall Prevention and Cognitive Decline Detection at Home

Ambient Assisted Living: Sensor Fusion เพื่อดูแลผู้สูงอายุที่บ้านด้วยระบบเซ็นเซอร์ผสานอัจฉริยะ

May 12, 2026 · 1 min read
Ambient Assisted Living: Sensor Fusion for Elderly Fall Prevention and Cognitive Decline Detection at Home

Ambient Assisted Living (AAL), a European Commission framework, uses unobtrusive technology to support elderly independence in daily life without sacrificing dignity or privacy. In Thailand — where 1 in 5 will be 65 or older by 2030 — AAL provides essential infrastructure for multi-generational families and elderly living alone.

Sensor Types and Their Specific Roles

PIR (Passive Infrared) sensors detect heat-based motion, establishing room occupancy status and building a daily activity baseline — which rooms are traversed at which times of day.

60GHz millimeter-wave radar (Seeed Studio MR60FDA1 or Aqara FP2) detects falls, monitors breathing rate, and tracks presence without cameras — ideal for bedrooms and bathrooms where privacy is paramount. Lab-tested fall detection accuracy exceeds 95%.

Door/window sensors capture detailed daily routines: what time the refrigerator is first opened (verifying breakfast was eaten), how many bathroom visits occur per hour (flagging polyuria or nocturia as potential diabetes signals).

A smart scale (Withings Body+) tracks daily weight. A rapid 2 kg gain over 3 days may indicate oedema from cardiac or renal issues, warranting clinical review.

Wearable integration (from post 23072) provides HRV, SpO2, and step count as physiological ground truth for fusion with environmental sensor data.

Sensor Fusion Logic in Home Assistant

Fusion combines multiple sensor streams to produce more reliable decisions than any single sensor alone.

Activity Score: count PIR triggers and door sensor events per hour, compared against a personal baseline (calculated from the first 14 days). If score drops more than 40% below baseline for 2 consecutive daytime hours → send LINE yellow-level alert.

Fall Alert: radar detects fall event + no movement for 60 seconds + no cancel button pressed on smart display → send LINE red-level alert with emergency contact numbers.

Cognitive Decline Flag: long-term pattern shifts — wake time drifting more than 2 hours from norm, refrigerator access frequency declining more than 30%, walking routes within the home changing significantly → generate a weekly report for physician review.

Privacy-First Design for Thai Cultural Values

Thai culture places profound importance on elder dignity. AAL systems must not deploy cameras in bedrooms or bathrooms — radar and PIR sensors provide sufficient safety intelligence without privacy compromise. All processing runs locally on Home Assistant with no data leaving to cloud servers, ensuring full PDPA (B.E. 2562) compliance.

Questions & answers

How do 60GHz radar and PIR differ, and should they be used together?
PIR detects coarse motion (requires walking) but cannot detect someone sitting still or lying after a fall. 60GHz radar detects presence via breathing even in a completely stationary person, and accurately identifies falls. Use both: PIR for cost-effective room-level occupancy, radar for high-accuracy fall detection in priority rooms.
Can AAL systems be installed in older homes without new wiring?
Entirely yes. Zigbee/Z-Wave radar and PIR sensors are battery-powered with no wiring needed. Door sensors use double-sided tape. Smart scales simply rest on the floor. The entire system is wireless, using Home Assistant as hub — no wall demolition or new electrical runs required.
How accurate is Home Assistant cognitive decline pattern flagging?
The system does not diagnose — it detects pattern changes that may warrant clinical attention. Accuracy depends on the 14–30 day baseline established during initial deployment. Research from Oregon Health and Science University found that activity pattern monitoring can detect cognitive decline 2–3 years before clinical diagnosis.

Related reading