MIT's autonomous block bots jump, roll and stack themselves
Out of all the cool-looking forms that robots can take – humanoid, dogs, fish, crocodiles, snakes, birds, or disembodied arms – a cube seems like a pretty boring choice. But MIT’s new take on the robotic block is smarter than it looks. These cute little cube bots can roll and tumble around, recognize each other, and stack themselves into whatever shape is needed.
Developed by MIT’s Computer Science and Artificial Intelligence Laboratory (CSAIL), the robots are made with flywheels inside them that spin at 20,000 rpm. When these brake suddenly, the angular momentum propels the block in that direction, letting them spin and flip their way around.
One lone block bot isn’t much use, but the idea is that they can swarm together to complete more interesting tasks. To do so, they have magnets embedded in each edge and face to let them snap together, and they have unique patterns on each face to let them recognize each other. MIT calls them “M-Blocks.”
“M stands for motion, magnet, and magic,” says Daniela Rus, CSAIL Director. “'Motion,' because the cubes can move by jumping. 'Magnet,' because the cubes can connect to other cubes using magnets, and once connected they can move together and connect to assemble structures. 'Magic,' because we don’t see any moving parts, and the cube appears to be driven by magic.”
These are actually the second iteration of M-Blocks. The original design, shown off way back in 2013, looked much the same but weren’t as smart. Now, the CSAIL team has developed algorithms that let them work together more effectively.
The researchers demonstrated that teamwork by having an autonomous swarm of 16 M-Blocks accomplish tasks like arranging themselves into a line, following arrows marked on other blocks, or moving towards a light source. When they get close enough together, they snap together magnetically. From that position they can even climb on top of one another by “throwing” themselves upwards. Once on top of other blocks, they can roll down the line, one face at a time.
The M-Blocks were fairly effective at achieving their goals, with about a 90-percent success rate. In the line experiment, for example, the blocks were programmed to look at their neighbors and figure out if they were connected to the right spot. If not, they would roll in a direction until they came to the end of the line.
While these tests only involved 16 of the blocks, the team says that the technology is scalable, with future swarms possibly numbering in the hundreds, thousands or even higher.
“The unique thing about our approach is that it’s inexpensive, robust, and potentially easier to scale to a million modules,” says John Romanishin, lead author on the study. “M-Blocks can move in a general way. Other robotic systems have much more complicated movement mechanisms that require many steps, but our system is more scalable.”
These swarming robot blocks can do more than just parlor tricks, too. The team says that eventually they could be put to work building temporary structures like staircases, to help people with inspections or disaster response. they could also end up in manufacturing, healthcare or gaming applications.
The research is due to be presented at IEEE’s International Conference on Intelligent Robots and Systems in November. The M-Blocks can be seen in action in the video below.