Untethered soft robots that swim by snapping

Battery-free locomotion powered by bistable snaps

Dates
2018
Collaborators
With Kristina Shea (ETH Zurich); and Osama R. Bilal, Chiara Daraio (Caltech)
A 3D-printed soft swimming robot with fins driven by a bistable element
A 3D-printed swimmer whose fins are driven by snapping bistable elements, with no battery, electronics, or tether.

Mobile robots typically depend on a tether or onboard power and control, each of which carries a cost: a tether constrains range and workspace, while batteries and electronics add mass, complexity, and points of failure. We developed small soft swimming robots that locomote with none of these: no battery, no electronics, no tether. Each is produced in a single multi-material 3D print, and each follows a trajectory specified entirely at design time.

Labeled design of the soft swimmer: outer shell, shape-memory polymer muscle, bistable element, floaters, blade pivot, and fins
The swimmer's anatomy: a shape-memory polymer muscle drives a bistable element, which pivots the fins; floaters and an outer shell set the buoyancy and constrain the motion. Everything is printed in one piece.

Mechanism of propulsion

The fins are actuated by a bistable element, a snap-through structure with two stable equilibria separated by an energy barrier. Crossing that barrier releases the stored elastic energy on a timescale far shorter than the loading timescale, producing the rapid, impulsive paddle stroke that drives propulsion.

The snap is triggered by a shape-memory polymer actuator: a printed curved strip programmed with a recovery temperature. When the surrounding water exceeds that threshold, the strip relaxes toward its programmed configuration and loads the bistable element. Shape-memory recovery is intrinsically slow, but using it only to trigger the snap decouples actuation rate from propulsion rate, so a slow thermal stimulus yields a fast mechanical stroke. The energy landscape is engineered to be asymmetric, so the forward transition is far more favorable than the reverse, rectifying the stroke into net directed motion.

Potential-energy and force curves of the bistable element versus distance, showing the asymmetric barrier and released energy, with programmed, activation, and activated states of the shape-memory muscle below
The mechanics. The bistable element's potential energy and force are asymmetric across the snap (left), so more energy is released going forward than back. Below: the shape-memory muscle in its programmed state, during thermal activation, and in its recovered state that triggers the snap.

A single muscle-and-element pair already produces directed swimming: as the muscle recovers and the element snaps, the fins sweep through one propulsive stroke and the robot advances.

Time sequence at t=0, 1.49, and 2.29 seconds of a single-stroke swimmer advancing through water
A single-stroke swimmer advancing over about two seconds as its one actuator fires.

Programming the trajectory

Multiple actuator-element pairs can be tuned to fire in sequence by setting their recovery thresholds: a thinner strip reaches temperature first, followed by a thicker one, yielding a coordinated train of strokes and longer travel. Because heading is governed by fin geometry and placement, the route is encoded in the morphology itself:

  • Two symmetric fins produce balanced thrust and straight-line motion.
  • Removing or offsetting a fin makes the stroke asymmetric, turning the robot through a predictable angle.
A swimmer with two actuator pairs and the time sequence of its sequential multi-stroke motion
A two-pair swimmer. Tuning each pair's activation threshold makes them fire in sequence, chaining multiple strokes for longer, programmable travel.

Composing fin layout with actuation timing lets us prescribe an entire path in advance. As a demonstration, we built a vessel that swims out, releases a cargo, and returns to its origin, with no feedback or control system of any kind.

A cargo-carrying swimmer with a shape-memory gripper and its sequence of swimming out, releasing the cargo, and returning
An autonomous cargo run. A shape-memory gripper holds a payload; the preprogrammed swimmer carries it out, releases it, and returns to its origin (t = 0 to 31 s), with no onboard control.

Significance

The result demonstrates that locomotion and rudimentary decision-making can be encoded in a material and its geometry rather than in electronics: here, the structure is the controller. This embodied-control approach suits low-cost, disposable, or otherwise hard-to-power platforms, such as untethered devices for open-water environmental or marine sensing, where an onboard battery and processor are impractical. The work appeared in PNAS.

A collaboration with Chiara Daraio's group at Caltech.

Related publications

  1. Chen T, Bilal OR, Shea K, Daraio C. Harnessing bistability for directional propulsion of soft, untethered robots. Proceedings of the National Academy of Sciences 115(22), 5698–5702 (2018). PDF