Gesture-Controlled Drone
A quadcopter platform controlled by hand gestures through a computer-vision pipeline — mapping detected poses to flight commands in real time.
Role
Sole designer
Team
Individual
Status
In Development
Timeline
2026 – Present
- ▹Hand-gesture recognition for flight commands
- ▹Computer-vision control loop
- ▹Custom quadcopter airframe assembly
- ▹Interactive on-site 3D CAD model
Problem Statement
Conventional drone remotes demand two-handed stick input and prior RC familiarity. This project explores a more direct interface: translating hand gestures from a camera into flight commands so control feels immediate and readable.
Design Goals
- Detect a small set of reliable hand gestures in real time
- Map gestures to takeoff, land, hover, and directional motion
- Keep the vision-to-command loop low-latency enough for stable flight
- Document the airframe, ESC, and flight-controller layout in CAD
System Architecture
Pipeline:
- Camera capture of the operator’s hands
- Computer-vision gesture classification
- Command mapping to flight setpoints
- Flight controller execution on the quadcopter
The mechanical assembly (frame, motors, ESC, GPS, and flight controller) is modeled in 3D for design review and portfolio presentation.
Hardware Design
Quadcopter airframe with four motor/propeller arms, onboard ESC, GPS module, and flight controller. The interactive GLB on this page mirrors the labeled CAD assembly.
Software Design
- Vision process for hand detection and gesture classification
- Control process that emits flight commands
- Safety defaults (hover / land) when confidence drops or gestures are ambiguous
Future Improvements
- Expand the gesture vocabulary with clearer separation between classes
- Add failsafe telemetry and battery-aware landing
- Outdoor bring-up notes and flight logs
Questions about this project? Email Caleb.