Wearable signals
HR, HRV/RMSSD, RRI, recovery, strain, duration, and activity labels provide physiological context.
A disability mobility case study showing why wearable systems need context for walking, skating, transport, recovery, and assistive movement.
Health AI Needs Mobility Data
A real world disability mobility case study showing how wearable data, route records, motion sensors, and source linked evidence reveal functional patterns that standard categories miss.
The core issue is functional mobility: For my documented pelvic, sacroiliac, and hip impairment, walking is active but ballistic, while controlled inline skating can preserve functional mobility. The evidence supports individualized review using medical records, accommodation history, wearable signals, route records, and source linked context.
The Evidence Model: Physiology, Motion, and Control
HandicapSkater is a within person mobility evidence record built from wearable data, motion sensors, route history, videos, notebooks, and functional testing.
The current corpus separates physiologic burden, mechanical motion exposure, and body coupling. HR, RMSSD, recovery context, HR per mile, distance per mean HR, duration tolerated, and RMSSD deltas belong in the physiologic layer. ACC dynamic RMS, jerk RMS, shock spike rate, and cumulative dynamic shock describe mechanical motion exposure at the sensor. Active controlled, active ballistic, passive passenger, and recovery baseline describe body coupling and control.
Inline skating and motorcycle riding can show higher raw motion while remaining active controlled mobility. Walking is active but ballistic. ParaTransit is passive imposed passenger exposure and must be reviewed as more than generic transportation.
The motorcycle with skates moment
One scene explains why this case matters: I leave a store wearing skates as a mobility aid, use a disabled parking space, get on a motorcycle, and ride away. Many people saw a stunt. The record shows an adaptation.
I began riding a motorcycle with skates because public transportation repeatedly refused access to the mobility aid I used to avoid high burden walking. I still had to get to work, appointments, groceries, court, and public life. When the accessible path was blocked, I built another one.
That moment exposes the platform problem: current systems cannot easily classify a disabled person shopping with skates used as a mobility aid, transferring onto a motorcycle, and using the same device for mobility, transportation, and access.
Beyond step counting
Most wearables count activity. This project asks the harder question: is the movement lower burden, higher burden, accessible, or forced?
For disability, more steps are not always better. The goal is functional mobility, not conformity to walking.
Route and video records
GPS maps and public videos show how mobility actually happens outside the clinic.
Source linked review
Evidence is organized so reviewers can separate source records, interpretation, metrics, limits, source quality, missing data, review boundaries, and claims.
Integrated Evidence Record
The current refined mobility science corpus contains 2,171 records. Body coupling context and physiologic efficiency fields exist for all records. RAG facets are expanded so records can be reviewed by activity, coupling context, physiologic burden, mechanical exposure, vehicle type, route context, audit status, and metric coverage.
The model separates physiologic burden, mechanical motion exposure, and body coupling. ACC, jerk, and shock describe mechanical exposure at the sensor. They do not automatically equal pain, physiologic burden, or functional burden.
2,171
Current refined corpus records
All records
Body coupling context and physiologic efficiency fields exist for all records
81
Records with ACC shock component metrics