TrAIT — Yuankun Huang

TrAIT

Context

TrAIT is a stroke rehabilitation system I built at UBC in collaboration with clinical researchers. Traditional upper-limb therapy is low-repetition and hard to sustain — conventional sessions produce roughly 32 therapeutic reaches; the clinical bar for motor relearning demands an order of magnitude more.

Decision

I built TrAIT in Unity with Leap Motion for precise 3D hand tracking and integrated the Challenge Point Framework for adaptive difficulty. The key decision was designing session telemetry that clinicians could actually trust in a peer-reviewed study — every reach timestamped, every workspace boundary calibrated to the individual patient, no data dropped on disconnection.

Outcome

The peer-reviewed study reported approximately 13% improvement in paretic arm function and roughly 31× more therapeutic reaches per session versus conventional care. Published in Neurorehabilitation and Neural Repair.

Gallery

Main Menu — Session entry point designed for clinical settings — large targets, minimal reading load — **Main Menu** — Session entry point designed for clinical settings — large targets, minimal reading load

Workspace Calibration — Each patient's reachable workspace is calibrated individually before every session — **Workspace Calibration** — Each patient's reachable workspace is calibrated individually before every session

Catch Task — Hand-tracked catching drives repeated reaching motions inside the calibrated workspace — **Catch Task** — Hand-tracked catching drives repeated reaching motions inside the calibrated workspace

Throw Task — Throwing variation keeps repetitions engaging while targeting the same motor patterns — **Throw Task** — Throwing variation keeps repetitions engaging while targeting the same motor patterns

Clinician Settings — Difficulty and task parameters adjustable per patient via the Challenge Point Framework — **Clinician Settings** — Difficulty and task parameters adjustable per patient via the Challenge Point Framework

Session Analytics — Timestamped reach telemetry surfaced for clinicians — the data backbone of the published study — **Session Analytics** — Timestamped reach telemetry surfaced for clinicians — the data backbone of the published study