The Berkeley Emergent Space Tensegrities (BEST) Lab at UC Berkeley, in collaboration with NASA Ames Research Center, is researching the use of tensegrity robots for planetary surface exploration missions. Our lab has designed several iterations of robots based on a spherical, 6-bar tensegrity structure, such as the one shown below.
I am curious if there are advantages to using larger tensegrity structures. A 12-bar tensegrity structure is the next-largest spherical, symmetric form. I suspect that a 12-bar structure will offer advantages in terms of actuation, impact, and payload capabilities. Funded by a NASA Space Technology Research Fellowship (NSTRF), I am leading the design and evaluation of new robots based on the 12-bar structure. I am beginning this research using rapid prototyping to learn more about the 12-bar form.
Unlike the 6-bar tensegrity, there are multiple geometric forms of 12-bar structures. I chose to focus on only Class I structures (one bar per node), and I learned of four spherical, symmetric 12-bar structures. The forms are named cube, octahedron, double-six, and rhombicuboctahedron. I created 3D models of each, as shown below.
Of these four structures, I chose three – the cube, octahedron, and double-six – to make an initial set of structural prototypes. I eliminated the rhombicuboctahedron because its rods contact one another in its interior. This contact would cause high localized stresses along the rods during impact and impede the rods’ motion during shape-shifting. I constructed the prototypes from wooden dowels and rubber bands, as shown in the figure below.
I experimented with these prototypes to gain design insights. I observed significant problems with the double-six structure, which has intersecting cables. The intersections made assembly of the prototype difficult and hindered its shape-shifting ability. In a full robot, the intersecting cables would cause fraying and snagging. I chose to postpone investigation of the double-six structure and focus on the cube and octahedron structures, which had promising properties without such limitations.
I iterated to create a final version of structural prototypes of the cube and octahedron, which are shown in the figure below. These prototypes were made by laser cutting the elastic members from sheets of silicon rubber, connecting these segments to form the lattice shell, and then attaching the rods to erect the structure. These structures are quick to assemble, easy to adjust, and mechanically robust. They will be used as the platform for actuated robots and drop-test analogs in my future work.
Thank you to the lab members who helped to create the prototypes: Cameron Bauer, Grant Emmendorfer, and Mariana Verdugo. Thank you also to Lee-Huang Chen, who developed the elastic lattice prototyping method used to create the final structures.