Despite differences in structure and living environment, several animal appendages (e.g., octopus arms,elephant trunks, etc.) follow a common pattern: the logarithmic spiral. This paper reports a new class of soft robots that morphologically replicate the spiral based on a common design principle across scales.This allows for speedy and inexpensive fabrication, while cables are used to provide a simple yet effective control mechanism. We further present a grasping strategy, inspired by the octopus, that can automatically adapt to a target object's shape. Through extensive tests, we illustrate the dexterity of SpiRobs and the ability to grasp objects that vary in size by more than two orders of magnitude and up to 260 times self-weight. Last, we demonstrate scalability via three additional variants: a miniaturized gripper, a 1-m-long manipulator mounted on a drone, and an array of SpiRobs that can tangle up various objects.
(A) A computer-aided design (CAD) model of a 25-cm-long SpiRob driven by two cables. First, the 2D discretized pattern is extruded to obtain a 3D solid (step①), which is then cut into a conical shape, and two holes are reserved for the cables to pass through (step ②).
(B) Comparison of the workspaces of different designs when holding the smallest graspable object: SR-, a SpiRob with lower deformation capability (obtained by decreasing the gaps while keeping the unit lengths fixed); CC-T, a tapered constant-curvature robot.
(C) Effect of elastic layer thickness on curling. If a thin layer is used, then the robot will collide with the base. Increasing the thickness can solve this problem, but it also means that a greater force is required to curl the robot into a spiral.
(D) SpiRobs with different taper angles: 5°, 10°, and 15°.
(E) Workspace: the smaller the taper angle, the larger the workspace. The workspace envelopes are calculated as in Note S2. The gray dots illustrate the reachable points (10 robot) generated by randomly sampling cable forces in simulation based on a joint-link model (Note S3).
(F) Theoretical predictions (Note S2) of the object size and weight that can be grasped with a maximum actuation force of 100 N: a larger taper angle means a larger weight for a fixed diameter.
(G) Images of a SpiRob (15) grasping a 5.6-mm-diameter cable, a 115-mm diameter tape, and a 10 kg weight.
(A) Snapshots of an octopus progressively uncurling its arm to reach, wrap, and grasp food. Inspired by this, we developed a strategy to replicate this motion by controlling the forces on two cables. Starting from a resting state (no force), we increase the force on the left cable, causing the robot to curl into a spiral from the tip (packing). Next, we increase the force on the right cable (F2) while keeping the one on the left (F1) unchanged, allowing the robot to uncurl from the base and reach out toward the object (reaching). Once in contact with the object, we slowly reduce the force on the left cable while maintaining the one on the right, enabling the robot to climb up the object’s surface and wrap around it (wrapping). Finally, we increase the force on the right cable to secure a firm grasp (grasping).
(B) SpiRobs grasping and moving various objects: grasping a raw egg, transporting a ping pong ball to a target behind the wall, and reaching through a crevice to grasp and retrieve a target object (more tests are shown in Figures S7 and S8).
(C) Graspable space for different-sized objects with the proposed strategy. Colored points represent the locations where objects (represented by different markers) can be grasped successfully. The envelope of the workspace is plotted as a dashed gray line.
(D) Graspable space for different-sized objects with the proposed strategy. Colored points represent the locations where objects (represented by different markers) can be grasped successfully. The envelope of the workspace is plotted as a dashed gray line.
(E) The robot automatically grasps objects of different geometries with the minimal information of object positions (p; a) obtained by a camera. Objects of different sizes are placed within the approximate area identified in (C).
The design principle based on the logarithmic spiral constitutes the main novelty in this study. Unlike in soft robotics, where the hardware is designed first, and the models are developed afterward, in our system, modeling (logarithmic spiral) comes first, and design/fabrication is a direct outcome of the model.
SpiRobs are bioinspired in both morphology and operation. A notable attribute is the scalability of the design principle (demon- strated with robots from centimeter to meter scale). They are capable of complex movements and feature formidable adapt- ability in handling objects that vary in size and shape with mini- mal actuation (2-3 cables). We further demonstrated three appli- cations: (1) a miniaturized robot that can handle fragile samples, (2) teleoperation on a portable platform (drone), and (3) a multi- arm gripper that functions via entanglement.
@misc{wang2024spirobslogarithmicspiralshapedrobots,
title = {SpiRobs: Logarithmic Spiral-shaped Robots for Versatile Grasping Across Scales},
author = {Zhanchi Wang and Nikolaos M. Freris and Xi Wei},
year = {2024},
eprint = {2303.09861},
archivePrefix= {arXiv},
primaryClass = {cs.RO},
url = {https://arxiv.org/abs/2303.09861}
}