Engineering, the milestones
Nothing is claimed until a milestone says so.
Six milestones stand between the specification and a kit anyone relies on, and the last one is the replacement test. Participants take part under an ethics protocol with accessible consent, and blind co researchers are paid authors.
The milestones
Measure, pass, status.
| Milestone | What is measured | What passing means | Status |
|---|---|---|---|
| G0, the bench | Every subsystem against its datasheet; the latency budget measured end to end; eye safety and battery certifications in hand. | All bench numbers meet the record's pass column; world to skin at 100 ms or less ahead and 150 ms or less at the ground, measured. | In preparation. The parts are chosen; the firmware is specified, not written. |
| G1, the controlled course, including signal recognition with blind wearers | Obstacle contacts, veering, and percentage of preferred walking speed against a mobility aid only baseline; each of the six signals played and named by blind wearers. | The doorframe at head height detected every trial, the glass door detected every trial, the step down before the leading foot; signals recognized reliably by the wearers who will depend on them. | Not started. Begins after G0 passes. |
| G2, real routes | Contacts, percentage of preferred walking speed, and walking speed on real indoor and outdoor routes; glass and drop off cases explicitly tested. | No missed glass door, no late step down, walking speed holding against the mobility aid only baseline. | Not started. |
| G3, daily life, the comfort milestone | Weeks of daily wear: comfort, wear time, skin condition, confidence, incident log. | Two hours of continuous wear with no skin injury, skin under a pod below 40 degrees, bands surviving 20 wash cycles, wearers choosing to keep wearing it. | Not started. |
| Test G, the replacement test | Proposed, to be confirmed with an orientation and mobility specialist: drop-offs of 5 cm or deeper across at least 500 approaches at normal walking speed; curbs and steps before the leading foot reaches the edge; measured end to end latency; supervised routes by blind mobility aid users with the mobility aid carried but not used; fault behavior; rain, bright sun, night, crowds, and stairs. | Zero missed drop-offs of 5 cm or deeper across at least 500 approaches. At least 99 percent of curbs and steps detected before the leading foot reaches the edge. Measured latency published and at or below the modeled figures. At least 10 blind mobility aid users complete supervised routes with a certified orientation and mobility specialist present. Every fault gives a distinct alert within a published time and the kit never fails silently. Rain, bright sun, night, crowds, and stairs tested and reported, including failures. | Not started. Proposed, to be confirmed with an orientation and mobility specialist. Only this milestone can change the line: keep traveling the way you travel now. |
| G4, the readiness review | Everything above, reviewed together. | Only after G0 to G3 pass are running mode and, separately, any assist frame work considered. | Not started. |