Soft Circuits to Control Soft Robots

Rubber circuits could help a robot fish rise and sink

3 min read
The toggle gripper holding a pair of pliers.
Photo: Harvard University

A human-operated, completely soft gripper reacts to the push of a soft button. This robot can pick up a variety of oddly-shaped or fragile objects, even after being struck by a mallet. Video: D.J. Preston, P. Rothemund, H.J. Jiang, M.P. Nemitz, J. Rawson, Z. Suo, and G.M. Whiteside

A team of scientists at Harvard University has built soft circuits that operate using just rubber and air, which they suggest could eliminate the last hard components from soft robots. They detailed their findings in the 25 March online edition of the journal Proceedings of the National Academy of Sciences.

Digital circuits symbolize each bit of data as either 1 or 0. The new soft devices symbolize data as different levels of air pressure in silicone rubber tubing—for example, 150 millibars represent 1, and zero millibars represent 0.

Electronic circuits are made up of logic gates—components that perform logic functions such as NOT, OR and AND that underlie digital computation. The new soft devices mimic logic gates using rubber valves that react to air flow in different ways. For example, the NOT logic function inverts the signals it receives, and when the soft NOT logic gate receives a high pressure indicating 1, its output is a low pressure indicating 0.

The soft OR and AND logic gates each require two inputs. The output of OR is 1 if either of its inputs is 1, otherwise it is 0. In contrast, the output of AND is 1 only when both its inputs are 1, otherwise it is 0.

The researchers used combinations of soft logic gates to create simple circuits that could, for example, demonstrate memory, retaining a 1 or 0 that could later get read as a source of pressure. Clock signals to help synchronize operations in these devices could be produced by soft pneumatic oscillators that cycle between high- and low-pressure states, they note.

The scientists developed a variety of devices controlled using soft logic. For example, they made a soft robot gripper claw clench or unclench by pressing a button. They also built a submersible robot equipped with a pressure sensor. It is designed to dive when it senses low pressure and surface when it senses high pressure or receives a signal from a control button.

A semi-autonomous submersible robot measures water pressure with a soft computer. It dives when it senses low pressure (near the surface) and rises when it senses high pressure (at depth). The robot also surfaces on command at the touch of an external button. Video: D.J. Preston, P. Rothemund, H.J. Jiang, M.P. Nemitz, J. Rawson, Z. Suo, and G.M. Whiteside

“The soft pneumatic digital logic demonstrated in our work enables robots and devices that exhibit memory and simple decision-making ability without any electronics or hard components,” says Daniel Preston, a mechanical engineer at Harvard University who was lead author of the paper based on the study.

The inspiration for soft logic came from increasing research into soft robots made of plastic and rubber. Whereas conventional robots are made of rigid parts that are vulnerable to bumps, scrapes, twists and falls, soft robots inspired by worms, starfish, and octopuses can resist many of the kinds of damage and squirm past many of the obstacles that stymie hard robots. Soft robots are also generally cheaper and simpler to make, comparatively lightweight, and safer for people to be around.

Scientists have found many ways to eliminate hard parts from soft robots. For example, instead of relying on rigid canisters of compressed air, researchers have explored powering soft robots using gas-generating chemical reactions or even explosions.

However, until now, soft robots have contained metal valves that open and close the channels of air that animate them, and electronics that instruct those valves when to move. “We saw an opportunity to directly integrate simple computational abilities, using the soft digital logic gates demonstrated in this work, into soft robots, to enable onboard decision-making without electronic control or hard valves,” Preston says.

Soft machinery “will be particularly useful for therapeutic and rehabilitation applications, where completely soft medical devices can reduce further injury or irritation [to patients],” Preston says. "With integrated soft digital logic, these medical devices can respond to stimuli from a patient—for example, pressing a soft button to turn the device on or off, or environmental stimuli [such as] a patient's temperature, or pressure points on a patient's body.”

Preston notes that these devices still require manual assembly of several parts. But “Moving forward,” he says, “we envision development of a simplified fabrication approach, using advanced manufacturing techniques, to allow large-scale integration of our logic gates in soft devices.” 

The Conversation (0)

The Future of Deep Learning Is Photonic

Computing with light could slash the energy needs of neural networks

10 min read

This computer rendering depicts the pattern on a photonic chip that the author and his colleagues have devised for performing neural-network calculations using light.

Alexander Sludds
DarkBlue1

Think of the many tasks to which computers are being applied that in the not-so-distant past required human intuition. Computers routinely identify objects in images, transcribe speech, translate between languages, diagnose medical conditions, play complex games, and drive cars.

The technique that has empowered these stunning developments is called deep learning, a term that refers to mathematical models known as artificial neural networks. Deep learning is a subfield of machine learning, a branch of computer science based on fitting complex models to data.

Keep Reading ↓ Show less