Robotics / Human–machine interaction

Gesture to motion.

A robotic hand that mirrors human gestures through computer vision and servo control.

Academic robotics projectWorking prototype

Conceptual illustration of hand landmarks and servo control.

01 / The challenge

The question behind the system.

Translate camera observations into useful physical motion across software, communication and mechanical interfaces.

A camera becomes the interface between a human hand and a physical robot. Python extracts hand landmarks, maps them to movement commands and sends them over UART to an Arduino Mega. Servos pull the tendons of a 3D-printed hand to reproduce finger movement.

02 / The approach

From architecture to implementation.

01
Detect hand landmarks with MediaPipe and process camera images with OpenCV and cvzone.
02
Map finger positions to servo commands and transfer them to the Arduino through a serial UART connection.
03
Integrate MG996R servos with a printed inMoov-based hand, tendon wires and elastic finger returns.

03 / Outcomes

What the work shows.

The prototype reproduces human hand gestures through the complete camera-to-actuator chain.
The repository includes the Python vision code, Arduino controller and original project report.

04 / Scope & next steps

Where the work stands.

Response depends on lighting, camera conditions, communication latency and mechanical friction or wire elasticity.
This is a gesture-control prototype; no quantified end-to-end latency or manipulation-accuracy benchmark is claimed.

The mechanical hand is adapted from the open-source inMoov3D humanoid model; the project integrates vision, control and the printed mechanism.

Field notes / Visual documentation