Read here about robots designed around the structure of a 6 strut tensegrity.

Super Ball Bot

Super Ball Bot is a name given to a 6 strut tensegrity robot being developed under the guidance of the NASA Ames Research Center in collaboration with graduate students from the University of California campuses at Santa Cruz, Berkeley, and San Diego as well as a student from the University of Ghent in Belgium. Dr. Alice Agogino's BEST Lab at University of California, Berkeley, is a partner in the research.

The Super Ball Bot attaches motors, batteries, sensors and elastomers to the struts and cables of a 6 strut tensegrity. These elements form an electronic control system that can loosen or tighten the tension of the cables. By varying which wires are loose and which are tense over time, the robot can collapse, expand or roll. The robot would suspend scientific instrumentation payloads in the middle of its body, and lower them to the ground to analyze surfaces and collect samples when necessary. Wireless communications systems in the robots will allow users to control the droids remotely.

Soft Robotics

The bot is a form of soft robotics. Soft robots push beyond the traditional approach of rigid robots, where forces magnify around joints and other common points of failure. First, the robots are collapsible and can be packed into compact shapes, a vital feature of a space-deployed robot since physical space is tight on such missions. The built-in elasticity of the 6 strut tensegrity framework allows forces to dissipate throughout the structure. NASA supports such development, since deployment is part of a robot's functional mission, and rigid robots require air bags and other cushioining, while a tensegrity has a potentially robust ability to absorb forces, and could land and move without extra mass. The control system is also distributed and redundant, so even if parts of a robot break, each part of its tensegrity structure is interdependent, so other parts can pick up the slack. One model has 4x redundancy and distribution, so even it a quarter of the motors fail, the structure should be able to continue rolling. The potential robustness of these robots could mean they could perform tasks traditional robots might not attempt, such as rolling off the edges of cliffs or down lava tubes.

Tensegrity also enables changing scale, or diameter. The fact that all the control systems of the robots fit into caps at the ends of each rod means robots of a variety of scales can be created based on oe working design, and a deployed robot could change its size dynamically by changing the lengths of its struts.

Deployment

A spaceship could drop Super Ball Bots, each covered by a heat shield to protect it from burning up in the atmosphere upon free fall. Note that many moons hvae a surface gravity a fraction of Earth's--Titan, for example, is one seventh--meaning that the robots' terminal velocity would be quite slow, 33 mph in the case of Titan. Planets, on the other hand, would result in greater velocity, due to thinner atmosphere and stronger gravity.

Control Systems

Controlling such round robots requires innovations in software and conceptual understanding of locomotion. They move with punctuated equilibrium by modifying strut lengths, causing their center of gravity to shift. To meet these challenges, a biologically inspired control system has been deployed. It features algorithms for controlling the robots that mimic central pattern generators, neural circuits in animals often vital to activities such as locomotion, chewing, breathing and digesting, which allow the robots to automatically roll in the same way that a biological creature breathes, without focusing on the task. Evolutionary algorithm development played a key role in developing these subroutines. These learning algorithms could be built-in to the control system, so different robots in different environments may develop different solutions to locomotion and payload dropping.

To help make tensegrity robots a reality, SunSpiral and his colleagues have released the open-source NASA tensegrity robotics toolkit, which is online for free and built on the Bullet Physics engine, a game physics simulator. See NASA Tensegrity Robotics Toolkit for more information.

Selected Publications

Rapid Prototyping Design anc Control of Tensegirty Soft Robot for Locomotion (video of experiments), July 2014.
"Presentation on Tensegrity Robots for Planetary Exploration," NASA Ames, March 21, 2013.
Jonathan Bruce, Ken Caluwaerts, Atil Iscen, Andrew P. Sabelhaus, and Vytas SunSpiral. Design and Evolution of a Modular Tensegrity Robot Platform. International Conference on Robotics and Automation (ICRA), May 2014.
Jonathan Bruce, Andrew P. Sabelhaus, Ken Caluwaerts, Alice M. Agogino, Vytas SunSpiral. SUPERball: Exploring Tensegrities for Planetary Probes. 12th International Symposium on Artificial Intelligence, Robotics, and Automation in Space (i-SAIRAS). June 2014.
Andrew P. Sabelhaus, Ken Caluwaerts, Jonathan Bruce, Alice M. Agogino, Vytas SunSpiral. SUPERball: Modular Hardware for a Mobile Tensegrity Robot. 6th World Conference on Structural Control and Monitoring (6WCSCM), Special Session on Tensegrity Syystems. June 2014.
Super Ball Bot - Structures for Planetary Landing and Exploration with Report and Presentation.

Links and References

For generic, non-robotic 6 strut tensegrities, see the 6 struts page.

Portal to Robotics
A series on robotics.
Robots: 6 strut, Biot, Robotics, Morphological Computation, Super Ball Bot
People: Agogino, Behzadipour, Hirai, Shai, Sterk, SunSpiral , Tensegriteam
Organizations: UC Berkeley
Software: Nasa's NTRT