Posts: 1
Joined: Sun Jun 30, 2019 10:17 pm

Complete beginner, Rubik's Cube robot Project

Sun Jun 30, 2019 10:28 pm


I am a complete beginner with only the coding experience of a typical AP Computer Science Java Course. I want to build a Rubik's Cube solving robot. Here is the inspiration for the project: ... ad3a2838cd

However, as I am extremely inexperienced in this aspect of the hobby/craft, I would appreciate any amount of help, especially the coding aspect of things.

In the article, the person has manually input the colors of the cube into a software he made, which then send the solving algorithm to the robot. Is there a way for me to use the Raspberry Pi Camera to scan all the faces and does the manual calculation on the board itself, as well as running the motors to execute the algorithm?

I assume the steps of the program will be:

1. raspberry pi cameras send all pictures of each face of the cube.
2. a program combines these pictures.
3. Using cube explorer's algorithm,, will return an algorithm, which the raspberry pi will execute by sending the steps to 6 steps motors, like the one in the website.

I have no idea how to execute any of these steps and would appreciate any help I can get.

User avatar
Posts: 148
Joined: Tue Feb 28, 2017 1:13 pm
Location: Sydney, Australia

Re: Complete beginner, Rubik's Cube robot Project

Wed Jul 03, 2019 2:08 pm

Your plan is valid and a good start.

One tip, Do your research, Do your research, Do your research, Do your research, as there are many articles on this subject.

Be sure to understand the rules if you want to compete, that is some time is allowed for scanning the cube before you start to solve it.

The mechanics will be your hardest aspect. Any half decent programmer can write a routine to solve the cube virtually on a PC screen, using graphics. It takes real skill to translate that to the real world. Moving physical objects (accelerating and decelerating their mass) at sub millisecond rates is no easy feat. Closed loop Stepper motors (you most likely choice) at these speeds do strange things.

Personally I would build a rough (and slow) prototype in say Lego first, and prove the basic software foundation, then having a basic working software, introduce the new hardware, then you go through a tuning and optimisation process of tweaking the design.

The Guinness competition saw the fastest robot solve a cube go from 8 seconds to just over 380mSec in 2 years. The world champion solves the cube in literally a blur of the eye. The mechanics are simply beautiful. See
..the only thing worse than a stupid question is a question not asked.

Return to “Automation, sensing and robotics”