FILE 01 / CLINICAL AND ENGINEERING CONTEXT
Foot drop frames the question. It does not validate the proposed answer.
Norysr is exploring whether one lower-leg motion sensor can recognize gait events reliably enough for a clinician-gated surface stimulation device to act on them. The clinical and engineering evidence behind that premise is under renewed source-by-source review.
This page describes the project question and boundaries. It does not provide medical advice or report a tested device.
01 / THE QUESTION
Can a wearable recognize the relevant part of an individual stride?
The project focuses on swing-phase foot drop in pediatric cerebral palsy. The v0.1 contract narrows the first engineering question to this: detect impending toe-off and heel strike causally, from one distal-shank inertial sensor, during supervised indoor level walking, and abstain when the signal, gait sequence or context is invalid.
That is an engineering question, not a conclusion. Norysr has no project dataset, no trained gait model, no prototype result, no clinical outcome and no validated trigger method.
02 / CURRENT CARE
The project does not replace established clinical assessment or care.
Orthoses, rehabilitation and functional electrical stimulation are clinical subjects with patient-specific indications, limits and risks. Norysr is not qualified to compare or recommend them while its source review is incomplete.
Any future stimulation settings, electrode placement, fitting workflow and participant criteria must be defined with licensed clinicians and appropriate safety review. The learned model never chooses any of them.
03 / THE ENGINEERING GAP
A useful prototype must expose its timing and failure behavior.
Timing
Measure the complete path from synchronized sensor input to supervisor decision, with freshness limits, not only model inference.
Comparator
Keep a transparent rule-based gait-event detector as the required comparator; a learned model is adopted only if it passes predeclared paired gates.
Failure behavior
Test dropped, stale, duplicated and out-of-order data, sensor faults, low confidence, reset and default-off behavior before any output exists.
Reporting
Report per-participant distributions, false and missed events and confidence intervals, never one best-case number.
04 / WHAT EXISTS NOW
A controlled specification and a tested software scaffold. No model, no hardware.
- A v0.1 build contract (August 2026) fixes the sensor, the model input and output, the deterministic supervisor and the ordered stage gates.
- A log-only software package implements schemas, causal filtering, a 32-frame ring buffer, deterministic replay, participant-split checks, sealed run bundles and a rule-based gait-event state machine, verified by 281 synthetic tests.
- No gait dataset has been licensed, no gait model has been trained, no hardware has been selected and no energized output exists.
- A separate computational side study on public EEG data has produced the workspace's first measured result; it is not a Norysr device result.
05 / NORYSR SCOPE
Keep the proposed contribution narrow enough to falsify.
- One calibrated 6-axis inertial sensor on the affected distal shank at 100 Hz; surface EMG is deferred to later evidence-gated ablations and is not a v0.1 input.
- The model reports three probabilities; a deterministic supervisor alone may convert a validated event into a bounded request, and every output stays disabled or simulated through the shadow stages.
- The provisional research cohort is ambulatory children aged 6 to 17 with unilateral spastic cerebral palsy and foot drop, GMFCS I to II, walking straight and level indoors under supervision. Clinical owners have not confirmed it.
- No stimulation envelope, participant protocol or regulatory path has been approved.