Mar 26 2019
Fixed to a balloon and immersed in a transparent water column, a soft robot dives and surfaces, and repeats the operation, similar to a fish chasing flies. Such a trick has been performed by soft robots before.
However, this one is different from a majority of soft robots and is developed and operated without using any electronic or hard components. Within the robot, the balloon is instructed by a soft, rubber computer on when to descend or ascend. This robot exclusively depends on soft digital logic, for the first time.
Over the past 10 years, soft robots have emerged into the metal-dominant realm of robotics. In assembly lines, grippers fabricated from rubbery silicone materials are already being utilized—for example, delicate vegetables and fruits like sausage links, celery, and tomatoes are handled by cushioned claws handle, or sweaters and bottles from crates are extracted by the grippers. In labs, live mice, slippery fish, and even insects can be picked up by the grippers, removing the requirement for additional human interaction.
Compared to hard robots, soft robots already need simpler control systems. The grippers are extremely compliant to such an extent that they cannot apply sufficient pressure to damage an item and without the necessity to calibrate pressure; here, a simple on-off switch can serve the purpose. However, so far, a majority of soft robots continue to depend on certain hardware—metal valves open and then close channels of air that control the rubbery arms and grippers, and these valves are instructed by a computer on when to move.
Now, using only air and rubber, researchers have developed a unique soft computer.
“We’re emulating the thought process of an electronic computer, using only soft materials and pneumatic signals, replacing electronics with pressurized air,” stated Daniel J. Preston, first author on a paper reported in PNAS and also a postdoctoral researcher who has teamed up with George Whitesides—the Woodford L. and Ann A. Flowers University Professor.
Computers utilize digital logic gates to make decisions. Digital logic gates are actually electronic circuits that not only receive messages (inputs) but also establish reactions (outputs) on the basis of their programming. The circuitry is not entirely different—when a physician hits a tendon below the kneecap (input), the nervous system is programmed to automatically jerk the leg (output).
The soft computer developed by Preston imitates this system utilizing pressurized air and silicone tubing. In order to obtain the least types of logic gates needed for intricate operations—in this case, AND, OR, and NOT—the soft valves were programmed to react to varied air pressures. For instance, in the case of the NOT logic gate, the output will be low pressure if the input is high pressure. Through these three logic gates, “you could replicate any behavior found on any electronic computer,” stated Preston.
For instance, the bobbing fish-like robot submerged in the water tank utilizes an environmental pressure sensor—that is, a modified NOT gate—to establish the kind of action to take. When the circuit detects low pressure at the tank’s top, the robot dives and then when the circuit detects high pressure at depth, it surfaces again. If an external soft button is pushed by someone, the robot can even surface on command.
Robots designed using only soft components offer a number of advantages. In industrial environments, for example, automobile factories, huge metal machines work with blind power and speed. A hard robot can cause permanent damage if a human gets in the way. However, if a soft robot bumps into a human, “you wouldn’t have to worry about injury or a catastrophic failure,” said Preston. This is because these robots can exert only that much force.
In addition to being safer, soft robots are usually affordable, lightweight, durable, resistant to corrosive materials and damage, and simpler to make. If intelligence is added, these robots can possibly be utilized for handling a lot more objects, besides tomatoes. For instance, a robot can alert a diver when the water pressure becomes extremely high; can perceive a user’s temperature and deliver a gentle squeeze to indicate a fever; and can push through debris following a natural disaster to help locate victims and provide assistance. Where electronics struggle, soft robots can venture easily—high radiative fields, such as those created in outer-space or following a nuclear malfunction, and also within magnetic resonance imaging (MRI) machines.
After flooding or a hurricane, a soft, hardy robot can possibly manage noxious air and hazardous terrain.
If it gets run over by a car, it just keeps going, which is something we don't have with hard robots.
Daniel J. Preston, Study First Author and Postdoctoral Researcher, MIT.
In this regard, Preston and coworkers are not the first teams to operate robots without the use of electronics. There are other research groups who have created microfluidic circuits, which can utilize both air and liquid to produce non-electronic logic gates. A soft octopus-shaped robot was aided by one specific microfluidic oscillator to flail all its eight arms.
However, microfluidic logic circuits usually depend on stiff materials like hard plastics or glass, and they employ such extremely thin channels that only tiny amounts of air can pass through at a time, slowing the motion of the robot. Preston’s channels, on the contrary, are bigger—around 1 mm in diameter—which allows relatively faster air flow speeds. Preston’s air-based grippers are capable of grasping an object in just a few seconds.
In addition, microfluidic circuits are known to be less energy efficient. These devices—even at rest—utilize a pneumatic resistor, which allows air to flow from the atmosphere to a pressure source or a vacuum in order to maintain stasis. When dormant, the circuits developed by Preston do not need energy input. Such conservation of energy can be vital in disaster or emergency situations, where the robots have to move far from a dependable energy source.
Furthermore, the rubber robots provide invisibility, which is an attractive possibility. Based on which material Preston chooses, he can easily create a robot that is index-matched to a particular substance. Hence, if he opts for a material that disguises in water, the robot would seem to be transparent when immersed.
In the days to come, he and his coworkers are hoping to develop autonomous robots that cannot be seen by the naked eye or even by sonar detection. “It's just a matter of choosing the right materials,” he stated.
For Preston, elastomers (or rubbers) serve as the right materials. Other fields pursue higher power with artificial intelligence and machine learning, while the Whitesides group turns away from the rising complication.
There’s a lot of capability there, but it's also good to take a step back and think about whether or not there's a simpler way to do things that gives you the same result, especially if it's not only simpler, it's also cheaper.
Daniel J. Preston, Study First Author and Postdoctoral Researcher, MIT.