Someday soon, your life may be saved by a weird-looking octopus, squid, or caterpillar – a squishy, form-changing, animal-like device that’s actually a soft robot.
Scientists are exploring the fluidity and deformable nature of animals like those mentioned above, as well as insects, starfish, and lizards. Their goal is to combine the maneuverability of these creatures with the autonomous nature of the hard-shelled robots we’re all familiar with—think R2-D2.
The ability of soft robots to climb onto textured surfaces and irregular shapes, crawl along wires and ropes, and burrow into complex, confined spaces will take them to places the hard robots of today can’t venture. In the biomedical field, they could assist in surgeries, while in search and rescue missions; they could crawl into hazardous situations to aid victims.
While a life-saving soft robot may not be a reality for several years, teams from various disciplines—computer science, organic chemistry, biomechanics, biomimetic robotics, flexible electronics, mechanical engineering, and materials development—are all making advances. Contributions to this field are also being made by neuroscience, polymer chemistry, control systems, and biomedical engineering.
Two groups at the forefront of this research are located in the Boston area. At Tufts University, professor Barry Trimmer heads one of the first groups to explore soft animal neuromechanics and how soft structures can be controlled through electrical motors.
At Harvard, a team of researchers led by professor George Whitesides, has developed a soft, silicone-based robot that looks much like an octopus. With a focus more on chemistry, the Whitesides Research Group is exploring elastomers, such as silicon polymer, and how the use of pneumatics—inflation and pressure—can change the shape of soft robots and power them.
According to Trimmer, soft robotics requires a new perspective for engineers. Typically, engineering theory deals with stiff materials and engineers have been trained to make sure what they build—whether it’s a bridge or a car—has minimal deformity under normal activity.
Soft materials, because they can change shape, are often considered to be problem that needs solving. “Soft material engineering is not taught much, and the engineering world needs to catch up fast,” Trimmer says. “Rarely are engineers encouraged to think about how they can build out of completely different materials.
“This would give them access to a whole new world of capabilities. Think of the proteins and sugars found in human bodies. These are amazing materials with fantastic properties that have never been used or exploited,” Trimmer says.
While this type of study won’t supersede current engineering, it will be an important part of the design of structures in the future. The promised applications are however some way off.
“There are still huge issues that need to be addressed,” Trimmer notes, “particularly in terms of developing a good framework and tools to support our theoretical approach.”
One area that needs to be advanced is soft material simulation tools. There is no means yet to build or model a device in a computer and simulate how it’s going to work as engineers do currently with structures such as aircraft.
Another challenge is the electronics that power the soft robots. “A rotary electric motor won’t work because it’s built of stiff materials and will limit the capability of a soft robot,” Trimmer says.
The Harvard team is using pressurized gas or liquid to power their robots, but it still has to use pumps and values and other equipment to drive the pressure. Trimmer’s group is exploring micro-coil shaped memory alloy wire, which can pull great force when current is passed through it. Trimmer and his team are also looking for a solution that is more like muscle, with electrically active polymers; the group is also using stem cells to grow muscles for their research.
The final major challenge is the control of movement of the robots. “Currently, engineering theory around controlling movement is for rigid systems,” Trimmer says. “This fails miserably when applied to deformable structures.”
Trimmer believes that advances in this area will come from computer science, morphological computation, and artificial intelligence as scientists explore the use of the material properties of structures to achieve control.
“Instead of having everything centralized in a computer, you actually use the body itself to accomplish many of the complicated tasks of control, much like how animals work,” Trimmer says.
To advance the study of soft robotics, Trimmer argues, it’s essential for engineers in different disciplines to be open to ideas and approaches from other fields. “In engineering, there is a tendency to be siloed. The people who build hard robots know mechanical engineering and work with control engineers. Because they have different backgrounds, these engineers tend to bolt the control system onto the robot, rather than develop an integrated system,” he says.
“The different disciplines need to be willing to get out of their comfort zone and collaborate with people in unrelated fields—even beyond engineering, such as physics, chemistry, and biology.”
We may see soft robotic toys and other advances like simple manipulator arms and gripper systems within the next five years. But, says Trimmer, it most likely will be another decade before completely soft devices will be put into use in more complex fields such as medicine.
Photo Credit: Tufts University