fof — wheelchair motion capture + therapy games

A portable motion capture system for manual wheelchairs, built as a clip-on kit for pediatric rehabilitation. The sensing stack measures wheel motion, chair orientation, and surroundings; therapy games (propulsion, turning, braking, precision) are applications on top of that telemetry.

The kit is non-invasive: magnets clip to spokes, the electronics box mounts to the frame, and nothing is permanently modified.

What we measure

Primitive	Source
Speed, distance, push count, direction	Hall sensors (A3144) + N35 magnets on spokes
Heading, tilt, smoothness (jerk), wheelies	MPU-6050 IMU (accel + gyro)
Obstacles, doorways, wall-following	HC-SR04 ultrasonic array
Start/stop, game input	Buttons / touch module / voice via companion app

Controller: Glyph-C6 (ESP32-C6, WiFi 6 + BLE 5, 20 GPIO).

Repo map

firmware/    PlatformIO project for Glyph-C6 — one lib/ module per concern
protocol/    Versioned JSON schemas: THE contract between firmware & apps
tools/       Python: record serial logs, replay algorithms offline, calibrate
apps/        Game dashboard (web client consuming protocol/)
hardware/    Mounts (3D print), wiring diagrams, BOM
docs/        Architecture, wiring, sensor notes, protocol spec, research
scripts/     setup / flash / ci helpers

Quickstart

./scripts/setup.sh          # install Python tools (+ PlatformIO check)
pytest tools/               # algorithm tests (no hardware needed)
./scripts/flash.sh          # build + flash firmware to Glyph-C6

Develop without hardware

Record one serial log from the kit, then develop algorithms offline:

python tools/record/capture_serial.py --port /dev/ttyACM0 --out data/session1.jsonl
python tools/replay/replay.py data/session1.jsonl

Status

Scaffold stage: firmware skeleton (encoders, IMU, ultrasonic, odometry, push state machine, JSON telemetry over Serial), Python algorithm twins with tests, protocol v1 schemas, CI for firmware build + tool tests. See docs/.

License

MIT — see LICENSE.