But the real breakthrough? It doesn’t just move - it adapts.
The Problem With Soft Robots
Soft robots are often modeled on nature’s most flexible movers - octopuses, worms, rays. But while they can bend and twist in ways rigid machines can’t, making them truly autonomous has been a challenge.
Batteries degrade quickly under constant mechanical stress. Control systems are often external. And piecing everything together side-by-side (a method called lateral integration) makes robots stiffer and bulkier, compromising their movement.
The team behind this study tackled that challenge by going vertical, literally.
Instead of laying out separate components, they stacked them, embedding flexible batteries and hybrid circuits directly into the robot’s soft body. This approach mirrors how biological tissues layer structure and function, and it freed up space, improved mechanical resilience, and kept the robot’s movements smooth and natural.
A Battery That Bends - and Lasts
It started with building the batteries.
The researchers developed flexible zinc–manganese dioxide (Zn-MnO2) batteries using a polyvinyl alcohol–graphene oxide gel and specially synthesized MnO2 cathodes. These were layered with laser-cut zinc and stainless steel foils, sealed in silicone, and - here’s the clever part - stabilized by the same magnetic fields used to move the robot.
Tests showed that when exposed to these fields, the batteries retained nearly twice as much capacity after 200 charge-discharge cycles: 57.3 % versus 31.3 % without the field.
Next came the body.
Inspired by the manta ray’s fluid motion, the robot was crafted from a soft silicone elastomer embedded with magnetic particles. The team magnetized the material in multiple directions (axially for steering, and in a wave-like pattern for up-and-down movement), enabling complex underwater motions. The flexible hybrid circuit, including an inertial measurement unit (IMU), was built to be just as compliant, with serpentine wiring patterns that could bend without snapping.
All of this was housed in a single, integrated unit that preserved the robot’s flexibility. Tests confirmed it maintained a low bending stiffness of just 0.348 newtons per millimeter, making it easy to steer with modest magnetic forces.
To control the robot, the team used two magnetic actuation setups: a mobile coil attached to a robotic arm for free movement and a flat electromagnet array for fine-tuned control of swimming trajectories.
Thinking on the Fly
But motion is only part of the story. This robot can also sense when something’s gone wrong and fix it.
Using real-time data from its onboard sensors, the robot feeds information into a digital twin: a software model that mirrors its position and motion in real time. If the robot gets pushed off course or bumps into something, it recognizes the disturbance and triggers corrective maneuvers using a nearby electromagnet array. In tests, the system kept the robot’s orientation stable and reduced angular drift, even during sudden perturbations.
It could steer through tight spaces, make U-turns, and recover from collisions - all without human intervention or a physical tether.
A Platform For Future Exploration
While the robot’s current runtime is limited to about 2.3 to 3.8 hours, the design opens doors for much more. The magnetic stabilization technique could potentially extend to lithium-ion batteries or other energy systems, and the vertical integration approach offers a flexible framework for building smarter, more autonomous soft robots.
Future versions could include chemical or ultrasonic sensors for environmental monitoring, or Hall effect sensors for even finer control. That makes this not just a one-off demonstration, but a platform for real-world applications, from underwater exploration to confined-space inspection.
Journal Reference
Li et al. (2025). Magnetic field–enhanced vertical integration enables embodied intelligence in untethered soft robots. Science Advances, 11(37). DOI:10.1126/sciadv.adv9572. https://www.science.org/doi/full/10.1126/sciadv.adv9572
Disclaimer: The views expressed here are those of the author expressed in their private capacity and do not necessarily represent the views of AZoM.com Limited T/A AZoNetwork the owner and operator of this website. This disclaimer forms part of the Terms and conditions of use of this website.