Turning repetitive ankle exercises into a maze game, because people actually do rehab when it's engaging.
Physical therapy for ankle injuries is clinically effective, but only if patients actually do it. Interviews with a physical therapist and a recovering patient surfaced the same story: at-home rehab exercises are repetitive, provide no feedback on correctness, and patients lose motivation quickly.
Without consistent exercise, recovery stalls and reinjury risk climbs. The real problem isn't medical; it's motivational.
GUARD (Gamified Ultimate Ankle Rehabilitation Device) is a foot-mounted platform that maps ankle motion to a custom maze navigation game. Dorsiflex to go forward, plantarflex to go back, abduct and adduct to turn.
An IMU sensor detects motion in real time. LEDs confirm when the correct range of motion is achieved. The maze can only be completed if all four prescribed movements are used, making the therapy the mechanic itself.
My work centred on the software interface layer, the part the patient actually sees. I built the front-end of the maze game in Python, including procedural maze generation that creates a fresh layout every session.
I also translated raw IMU data into in-game directional commands, calibrating the sensor readings so foot movements map intuitively to character motion without lag or jitter.
On the electronics side, I handled the integration between the IMU, Arduino firmware, and LED feedback system, ensuring the threshold logic triggers correctly when each range of motion is achieved.
Physical prototype: 3D-printed footplate with IMU
Python maze game interface