Arabnet Arabnet

Phonebloks and Project Ara got our attention to modular design, an approach that subdivides a system into smaller parts - or modules - that can be independently created and then used in different systems to drive multiple functionalities (definition from wikipedia). But what are some of the other applications of this concept?

At the Massachusetts Institute of Technology (MIT), researchers were able to build cube robots, which they called M-Blocks, that can climb over or around one another, jump, roll on the ground, or move while suspended upside down from metallic surfaces. And when we say cubes, we mean just cubes; these small devices don’t have any external moving parts or mechanical limbs to help them move. See for yourself.

Source:http://firewireblog.com/2013/10/10/mit-researchers-create-m-blocks-modular-self-assembling-robots/

How is this possible?

The cubes use angular momentum to propel themselves and magnets to stick to one another. Inside each cube you’ll find a motor, a motor controller, a flywheel, a breaking mechanism, a radio, a battery, and magnets. The flywheel is capable of spinning at about twenty thousand revolutions per minute. When the flywheel is braked, the angular momentum is transferred onto the cube which get pushed, or thrown, forward.

“What’s unique is that the spinning mass is inside the robot, which means that the robot doesn't need to be in a certain position in order for the force to be acted on the robot. This allows for a lot more types of motion with only one actuator.”

The magnets are placed on the edges and on every face of the cube in a way that allows any two cubes to attach to each other. The edge magnets are cylindrical, mounted like rolling pins, and can freely rotate. So when two cubes are near to each other, the north poles naturally align with the south. This means that any face of any cube can attach to any other. The faces of a cube contain four more pairs of symmetrically arranged smaller magnets that help snap a moving cube into place when it lands.

The idea was to design self assembling and self re-configuring robot systems that have the ability to change their geometry. Unlike non modular robots, which, having a fixed architecture perform specific tasks well but fail to adapt to new environments, modular robots are much more versatile.

“If we do not know ahead of time what the robot will have to do and when it will have to do it is better to consider making modular robots that can attain whatever shape is needed for the manipulation, navigation, or sensing needs of the task.”

During conception, the cubes were not intended to fly, but it turned out that the cubes had enough momentum inside them that, when maximum energy was applied, they threw themselves through the air. The magnets serve to bring them into alignment when they land.

At the moment, the cubes are controlled on a very basic, almost literal, level; meaning that every action, or step, is inputted into a computer and then transferred over a wireless link to the modules which simply execute what they are told. The vision for the future however is different, and much more exciting. The researches hope to put an algorithm on the modules directly that would enable them to make decisions for themselves. That means that commands will be issued to a group of cubes on a very high level, perhaps only describing the end goals of a task. It is them up for them to decide how to go about the job.

Possible applications for such a technology are limitless and fascinating to ponder. As a first step, cube armies could perhaps be deployed to assist workers in repairing structures. Additionally, more technology could be added onto the cubes: different types of sensors, cameras, microphones, lights, and so on.

A more ambitious outlook sees these devices miniaturized even further, so small that they form hordes of swarming micro-bots that would look and act like liquid steel.