[ 00.03 ]  PROJECT 42.7370°N / 84.4839°W

Hand-controlled drone.

A gesture-driven quadcopter prototype — IMU input on a wearable, onboard firmware parsing pose into flight commands.

IN PROGRESS
YEAR / 2026 ROLE / Hardware · firmware TOOLS / KiCad, C

A gesture-controlled quadcopter prototype: the hand wears an IMU, and the drone takes the pose as a control input.

I wanted a project that would force me across the full hardware-to-firmware stack rather than stopping at the PCB. A gesture-driven drone made the decisions unavoidable — what does the sensor give you, how do you turn it into a stable command stream, and how much of that work belongs on which end of a wireless link.

The system has two sides. The wearable carries an IMU, a microcontroller, and a wireless module; it samples orientation, converts pose to commands, and transmits. The drone side parses incoming packets and forwards them to the flight controller. Design decisions worth naming:

  • IMU on the glove, sampled at a rate matched to the control loop
  • Complementary filter fusing accelerometer tilt with gyro rate into a pose estimate stable enough to control against
  • Wireless link; packet rate sized for the control loop, not for the sensor’s full sample rate
  • Firmware deadband and slew limits so small hand tremor doesn’t map to erratic drone input
  • Fail-safe: if the link drops for more than N ms, the drone commands neutral throttle and holds

Where it stands: the wearable is assembled, the filter is behaving, the link is up, and the drone side is receiving. I’m still tuning the mapping from pose to flight command — tilt-for-pitch is obvious, yaw is harder, and altitude from a glove is genuinely unresolved. Next revision is a second wearable with a thumb-actuated throttle and a more conservative default mapping.