 |
The world record Rubik's cube solving machine of a group of students. Photo: NVCC . |
A group of Purdue University students recently set a new Guinness World Record with a robot they designed, solving a Rubik's Cube in just 0.103 seconds. That's three times faster than the previous record set by another robot.
The record wasn't achieved by a faster-moving robot. The students combined a high-speed but low-resolution camera system, a Rubik's Cube customized for durability, and a special solving technique popular among speed-solving professionals.
Key differences from competitors
The Rubik's Cube Robot Race began in 2014, when a robot built from a Lego Mindstorms kit and powered by a Samsung Galaxy S4 solved the cube in just 3.253 seconds. In May 2024, engineers at Mitsubishi Electric in Japan claimed to hold the world record with a robot solving the cube in 0.305 seconds.
To get the robot’s solving time down to under half a second, the team ditched Lego components and replaced them with optimized parts like industrial motors. But to reach the 0.103-second mark, the Purdue team tweaked the speed at which their robot could “see” the Rubik’s Cube’s shuffle.
The speed solver can observe the cube before the timer starts. However, the timer will also take into account the time it takes the robot to identify where each colored square is on the faces.
The students used two high-speed Flir machine vision cameras, with a resolution of just 720x540 pixels, placed at opposite corners of the cube. Each camera could simultaneously observe three faces of the cube in a single shot lasting just 10 microseconds.
 |
Robot's extremely fast color recognition technology. Photo: NVCC. |
Conventional cameras still need time to process the data from the sensor and turn it into a digital image. However, Purdubik's Cube, as the students' robot is called, uses a custom image recognition system that skips the image processing step entirely.
The system focuses on a very small area (128x124 pixels) in the frame captured by each camera, reducing the amount of data that needs to be processed. The raw sensor data is sent directly to a high-speed color detection system that uses RGB values from even smaller sample areas to determine colors faster than both conventional methods and AI.
Although the method is unstable, it can meet the team's prerequisite. "Even if the accuracy is only 90%, that's still good enough. What we really need is speed," said Patrohay, a student in the team.
Optimize the whole system
Patrohay said that each previous record-breaking robot usually improved on one prominent element. The MIT student team's robot (2018) focused on using high-performance industrial hardware. The Mitsubishi Electric team chose a specialized electric motor, designed to optimize the rotation of each face of the Rubik's Cube.
Meanwhile, the Purdue team chose off-the-shelf software to optimize the entire system, from the camera, image processing, hardware, to the solving algorithm. They used Elias Frantar's Rob-Twophase, a Rubik's Cube solving algorithm specifically for robots, allowing them to take advantage of special capabilities such as rotating two Rubik's faces at the same time.
The team also utilized a technique that allowed them to start rotating one side of the cube before completing the rotation of the other side perpendicular to it. This method saves a lot of time, but it also risks damaging or even shattering the cube if the timing is wrong or if too much force is applied. Therefore, the students also had to customize the cube to withstand the force and operate smoothly with this technique.
According to the World Rubik's Association (WCA) rules, participants can customize their own Rubik's Cube, as long as it can still rotate and function as a standard cube, with 9 colored squares on each face, and 6 faces with 6 different colors. Players can use materials other than plastic, but the colored parts must have the same surface texture.
 |
6 motors like this will be attached to the remaining sides, taking on the task of rotating the Rubik's cube. Photo: NVCC. |
To increase durability, the Purdue team upgraded the cube’s internal structure with a custom 3D-printed version made from a stronger SLS nylon plastic. Lubrication and increased tension also help reduce overshoot and increase control.
Purdubik's Cube uses six motors attached to a metal shaft centered on each face of the cube. After testing several different methods, the team settled on a trapezoidal motion system that allows the robot to precisely align each face when it comes to a stop.
Patrohay believes Purdubik could have broken his own record if the cube had been made from a more durable material other than plastic. “If you made a dedicated cube entirely from carbon fiber composite, I think it would have been able to handle higher speeds. You could have continued to reduce the time,” he said.
Source: https://znews.vn/ben-trong-robot-giai-rubik-nhanh-nhat-the-gioi-post1557575.html
Comment (0)