Researchers create reconfigurable robots with muscle-inspired HEXEL modules

Researchers have created flexible, biologically inspired modules that connect magnetically to form reconfigurable, high-speed robots.

The HEXEL modules developed by the Robotic Materials Department at the Max-Planck Institute for Intelligent Systems (MPI-IS) can contract linearly like mammalian muscles.

Experiments showed these modules can create robots with capabilities such as rolling, jumping, and crawling.

“Integrated artificial muscles into hexagonal exoskeletons that are embedded with magnets, allowing for quick mechanical and electrical connections,” said reseachers, in a statement.

Versatile soft actuators

Modern robots excel in specialized tasks but struggle with versatility in unpredictable environments while managing costs. Modular robots offer a solution, enabling quick rearrangement and adaptability, reusing parts to enhance sustainability and robustness.

Soft actuators, inspired by natural muscle, provide lightweight, conformal solutions with high adaptability, safety, energy efficiency, and durability. However, they often lack the blend of high-speed and high-strain actuation needed for agile, multi-functional robots.

Pneumatic actuators, though effective, face reconfiguration and efficiency challenges. Shape memory alloys (SMAs) and dielectric elastomer actuators (DEAs) offer untethered operation but are limited by slow response or manufacturing difficulties.

Soft electrostatic actuators with solid-liquid dielectrics provide fast, versatile actuation and are promising for modular, reconfigurable robots.

In the study, the team at MPI-IS built a robotic module with a solid hexagonal exoskeleton and soft actuator pouches filled with liquid dielectric material at the connections, inspired by the way bones translate forces generated by muscles.

The actuator pouches’ electrodes zipped together in response to a voltage applied across them, pushing the liquid dielectric between them in the direction of their vertices. The module was forced to contract rapidly due to the increase in hydraulic pressure, which triggered actuation.

Reconfigurable robotics

Researchers claim that individual HEXEL devices may power robots that can crawl through small pipelines and jump more than four times their own height.

In order to facilitate simple and reversible connections between HEXEL modules, the team additionally incorporated magnets into the stiff plates and electrodes.

They demonstrated that magnetically connected HEXEL modules could generate more power and a wider range of movements as compared to solo units. A strong rolling motion might be produced by successively activating circularly connected modules to tilt quickly in the same direction, enabling the robot to move over pebbles and through the sand.

Likewise, an artificial muscle with enhanced contractile capacities was produced by arranging a series of vertically connected HEXEL modules. In order to power untethered robots, researchers also developed a magnetic power source that could be snapped onto HEXEL modules.

“Combining soft and rigid components in this way enables high strokes and high speeds. By connecting several modules, we can create new robot geometries and repurpose them for changing needs,” said Ellen Rumley, a visiting researcher from the University of Colorado Boulder, in a statement.

According to researchers, developing robots with reconfigurable capabilities is a sustainable design choice. Instead of purchasing multiple robots for various tasks, we can use the same components to build different robots for diverse purposes.

“Robots made from reconfigurable modules could be rearranged on demand to provide more versatility than specialized systems, which could be beneficial in resource-limited environments”, said Zachary Yoder, a Ph.D. student working in the Robotic Materials Department, a co-first author of the publication, in a statement .

The authors propose that advancements in attachment methods could lead to self-assembling reconfigurable HEXEL robots.

The details of the team’s work were published in the journal Science Robotics.