Teaching a Robot to Build Patterns

A robot that picks, moves, and places particles to build structured patterns

Key insight

A robot that builds patterns from individual particles sounds simple — until it has to do it reliably.

Each step must work every time:

  • find the particle
  • pick it without failure
  • move precisely
  • place it exactly where it belongs

This project turns that sequence into a repeatable, autonomous loop.


System

  • Robotic arm with vacuum suction end-effector
  • Python control pipeline for motion and sequencing
  • Color-based particle detection
  • Goal: assemble predefined patterns from individual particles

What I did

Built a full pick-and-place loop:

  • move to particle location
  • detect particle (color-based)
  • activate suction and pick
  • move to target position
  • release suction and place

On top of this:

  • generated structured patterns (e.g., Yin–Yang)
  • tested with particles of different shapes and sizes

Methods & tools

  • Language: Python
  • Control: sequential command pipeline
  • Gripping: vacuum suction system
  • Perception: color-based detection
  • Task: discrete particle manipulation and placement

Key results

1. Reliable manipulation

  • Consistent pick-and-place across repeated trials
  • Accurate positioning at target locations

2. Pattern formation

  • Successfully assembled predefined designs
  • Structure emerges from repeated local actions

3. Robust behavior

  • Works with both spherical and irregular particles
  • Handles variation without changing the pipeline

Takeaway

Complex structure does not require complex logic.

With a simple perception–action loop, a robot can reliably pick, move, and place particles to build organized patterns.


Media


▶️ View more videos →