SoftSnap: Rapid Prototyping of Untethered Soft Robots Using Snap-Together Modules

Abstract

Soft robots offer adaptability and safe interaction with complex environments. Rapid prototyping kits that allow soft robots to be assembled easily will allow different geometries to be explored quickly to suit different environments or to mimic the motion of biological organisms. We introduce SoftSnap modules: snap-together components that enable the rapid assembly of a class of untethered soft robots. Each SoftSnap module includes embedded computation, motor-driven string actuation, and a flexible thermoplastic polyurethane (TPU) printed structure capable of deforming into various shapes based on the string configuration. These modules can be easily connected with other SoftSnap modules or customizable connectors. We demonstrate the versatility of the SoftSnap system through four configurations: a starfish-like robot, a brittle star robot, a snake robot, a 3D gripper, and a ring-shaped robot. These configurations highlight the ease of assembly, adaptability, and functional diversity of the SoftSnap modules. The SoftSnap modular system offers a scalable, snap-together approach to simplifying soft robot prototyping, making it easier for researchers to explore untethered soft robotic systems rapidly.

Figures (5)

• Figure 1: Overview of the prototyping process. The first row demonstrates that combining $n$ SoftSnap modules with a customized connector can construct a soft robot. The second row illustrates a gripper robot made from four SoftSnap modules arranged in a square and connected in the center. The third row depicts a brittle star-like robot created with five SoftSnap modules in a pentagonal configuration, also connected centrally.
• Figure 2: Examples of the cable-driven SoftSnap modular system applied to bio-inspired robotic designs. The image illustrates three robotic motions inspired by nature. The right images show control signals for five modules, highlighted in blue, red, yellow, orange, and green. (a) The starfish-like robot replicates the turning-over behavior observed in real starfish during 10.5 s. At t = 0 s, the robot leans on a purple block shaped to resemble a starfish. By t = 1.75 s, the robot achieves a similar posture to a real starfish, with its top leg raised. It continues to receive control signals until it falls due to gravity at t = 3.25 s. The robot vividly emulates starfish motion until t = 11.25 s, after which the control is switched off, and the robot relaxes, resembling a starfish resting on the seafloor. (b) The brittle star-inspired robot demonstrates locomotion through periodic movements with a cycle of 1 s. Initially, the robot lies flat with its legs extended. A large step control signal is applied for 0.5 s, causing the legs to contract rapidly, followed by a second control signal for the remaining 0.5 s to release the actuation and contract the other legs. This stretch and shrink motion generates momentum, showing clear displacement from its starting position. (c) The snake-like robot performs concertina locomotion by sequentially deforming its modules. In this demonstration, the robot exhibits a periodic motion with a cycle of 1.1 s. It starts in a straight position, contracts its body, achieves full contraction, begins to unwind, and progressively returns to its initial straight state from head to tail. This is achieved by periodically alternating the step signals in different modules. The picture on the left shows that at t=0 s, the robot is in its initial state; at t=1 second, it begins to contract; at t=2 s, it is fully contracted; and by t=3 s, it is midway through returning to its original state. Credits for the three real animal images: SeaAnimal4k (Star fish 4k Amazing Starfish in Undersea Ultra Hd, https://www.youtube.com/watch?v=MWlKfaxROd0), keepondiving3140 (Black brittle star Small Giftun Red Sea Egypt, https://www.youtube.com/watch?v=cX9c43sF3vo), and SnakeDiscovery (How Snakes Move! (They don't just slither!), https://www.youtube.com/watch?v=7-AKPFiIEEw&t=30s).
• Figure 3: Demonstrations of the cable-driven modular system applied to manipulation tasks.(a) The gripper robot effectively grasps objects of different shapes, such as a ball and a cuboid. In a(i), we illustrate the five stages of ball grasping. In a(ii), a(iii), and a(iv), we demonstrate grasping with three body shapes: one cuboid and two irregular forms, all controlled by the same signal for the four modules. (b) A four-module ring configuration is used for caging-type object transport, showcasing the ability to transport a large ball. At t = 0 s, the ring cage is in its initial state.
• Figure 4: SoftSnap module design and motor module details.(a) Three components of the actuatable module: passive skeleton, 3-in-1 motor module, and two-directional connector. The passive skeleton is linked to the 3-in-1 motor module via a winch. A string extends from the motor, passes over the winch, threads through the holes on the passive skeleton, and continues to the far end of the passive skeleton. (b) The exploded-view drawing of the 3-in-1 motor module components which satisfied the three functionalities: Power, Control, and Actuation. (c) The PCB's mother board and daughter board are displayed on the left. The system function flow is present on the right.
• Figure 5: Five different threading methods and their resulting deformations after shortening the strings to the same length in both simulations and real experiments. (a) The left end of the skeleton, attached to the body case, is fixed to the ground. The blue line illustrates the single string that runs from the motor to the skeleton and reaches its far end. The red line represents the deformed shape of the skeleton along its longitudinal axis. (b) Example of a '' w" (curve-5) shape configuration, comparing seven steps of shortening the strings evenly in experiments and simulations. (c) Displays all the configurations with the 'Curve-3' threading method and shows the workspace of the middle of the skeleton from 0 mm to 80 mm string contract length; the contraction process is associated with the color bar. (d1) and (d2) The average RMSE is calculated with the RMSE of the 12 center points (marked in pink) of each rod on the skeleton. (d1) Illustrates the average RMSE across five experiment threading patterns across seven testing steps. Within the experiment range, the real robot shows a good match with the simulation. (d2) Shows the RMSE of five different experiment threading patterns over steps. The average RMSE, with a maximum of about 6 mm, is relatively small compared to the skeleton's length of 201 mm. The error has an increment trend with the string length contraction.