NEWS
The bistability of fidget poppers inspires enhanced walking and grasping capabilities for soft robots.
Written by Jennie Morton
SOFT ROBOTS ARE PRIZED for their springy, bendable features. Unlike their rigid counterparts, these machines are designed for flexibility. But programming their pliable movements is a challenge because there are infinite variables. A solution lies in the elasticity of a fidget toy.
Sensory poppers have the ability to flex between two stable states. Research led by Andres Arrieta, a professor of mechanical engineering at Purdue University, has demonstrated how to strategically activate the material to enable functions like walking and gripping.
“The goal is to create soft robots that have significant movement capabilities but with little to no computing,” Arrieta explained. “Rather than rely on a neural network, the robots are steered with a local input that is essentially married to their morphology.”
This is accomplished without electricity or a central processor. The end goal is to create robots that don’t depend on electronics like onboard computers, batteries, or sensors.
This robot exhibits mechanical intelligence, where the ability to hold different apertures is accomplished by actuating domes embedded in its arms.
Photo: Andres Arrieta/Purdue University
The Basics of Bistability
Beyond their ability to regulate busy hands, poppers exhibit bistability. This is when a material can hold two distinct states, like the on/off position of a light switch, retractable pen, or mousetrap. Nature offers examples in the form of earwig wings or the leaves of the Venus flytrap, Arrieta added.
Made from thermoplastic polyurethane, the bistability of poppers offers two valuable qualities. First, it takes little force to move the domes between their two positions. They can also have metastability, which is when a material reverts back to a base state after a specific interval. Arrieta’s team wondered—could these physical properties be harnessed as a form of mechanical programming?
To test the idea, the team integrated domes between the articulating arms of a gripper robot. The domes are pneumatically actuated with compressed air, curling into a progressive curvature like a finger being crooked.
Different apertures are predetermined since the distance and timing fluctuations between each dome’s two states are known variables. The arms then successfully maintain their grasps around an object until the domes return to their starting positions.
Thermoplastic polyurethane is also beneficial because it is widely available, easy to 3D print, and affordable. But other stretchy polymers like silicon have bistability that could be leveraged as well.
“It’s the geometry that dictates functionality, not the material itself,” Arrieta said.
“It’s the geometry that dictates functionality, not the material itself.”
—Andres Arrieta, Rising Star Associate Professor at Purdue University
Outside the Binary
Robots based on biomimicry open new possibilities for use cases. For example, machines with fewer electronics are better suited for harsh, corrosive environments like mining operations and nuclear reactors. They could even perform tasks in the ocean and space.
“We need to develop robots that function more like organisms with behaviors that don’t require active input from a central nervous system,” Arrieta emphasized. “For example, humans don’t consciously think about the way we walk. The brain sends a sign for a central pattern generator that our legs follow. This circuit is possible because there are neurons locally connected to the muscles. Only a large pattern disruptor like ice will require intervention.”
It’s this type of “local computation loop” that Arrieta’s team is tapping for robotic movement. There are no commands from a central processing system—just the malleable domes distributed throughout the structure acting as inputs.
Because these soft robots aren’t using electronics, they are better able to withstand damage. These pins were unable to degrade the system’s overall functionality, demonstrating how they could maintain performance in a rough environment.
Source: Andres Arrieta/Purdue University
“Our concept is based on operational set points, where we program the dynamics of the robot to perform a particular task based on the geometry of its structure,” Arrieta explained. “Rather than using 1s and 0s for computing, we draw from control theory and structure’s inherent stability to let it compute and respond in a desired way.”
This design simplicity is also important because soft robots are meant to have “infinite degrees of freedoms,” so it’s a challenge to predict and then program all the ways they can move, Arrieta added. Using software also requires chips and sensory systems, all of which are vulnerable to failures.
The team did further testing with a walker robot, achieving dynamic movement with a combination of metastable and bistable domes. They even deliberately damaged the arms of the gripper robot, demonstrating that direct penetrations don’t interfere with its performance. These findings illustrate how mechanical intelligence in soft robots simplifies and strengthens their control.
Jennie Morton is an engineering and construction writer based in Iowa.
© 2026 The American Society of Mechanical Engineers. All rights reserved.