Rubik's Cube Robot

Personal Project

Fall 2016 - Fall 2017

Starting as a hackathon idea, the goal was simple. Solve a Rubik's cube automatically. Though it started as purely developing the algorithm, the engineer in me quickly became determined to build a robot to actually physically solve the cube.

As one of my earliest 'engineering' projects, I went through many many iterations, learning a lot along the way. Though the final product is not quite polished, the journey was of such great importance since it was effectively my introduction to robotics. I enjoyed the entire process, from the physical design of the 3D printed structure and mechanical arms that turn the cube, to the computer vision algorithm to detect the orientation of the cube. I learned so much from this project. Maybe someday I will come around to redo it right (without the duct tape).

​

​(To be clear, the robot does work for most extent. Two cameras capture the cube's orientation, the solving algorithm runs in under 1 millisecond, and the motors can solve the cube in under 2 seconds consistently. Just that there are aspects that are somewhat finicky, such as the camera capture dependence on lighting (even with the LEDs), and unpredictable motor hiccups that jam the cube)

​