Tensegrity robots

The term tensegrity, introduced by architect Buckminster Fuller, originates from the conjunction of the words “tension” and “integrity”. Tensegrity structures are a particular type of prestressable structures consisting of compression elements connected through tension elements. Traditionally, a mechanical structure (e.g., a building or an aircraft wing) is made of orthogonal rigid elements, which do not usually yield the minimal mass design for a given set of stiffness properties, and thus have to be overdesigned to deal with loads in many different directions. On the contrary, the most important characteristic of tensegrity structures is the absence (or, in practice, the negligible amount) of bending and shear forces, so that their structural elements are exposed only to unidirectional loads, and can achieve a very high strength with small mass. As a consequence, it would be possible to deploy lighter robots to be used in heavy-duty industrial applications as compared to the state of the art, as the maximum payload-to-weight ratio for a tensegrity structure can be much higher than that of a conventional manipulator.

One might wonder why, in spite of the huge potential of tensegrity manipulators, we still lack a general motion planning and control framework for them. In our opinion, this is due to two main problems. The first is the complexity of the system dynamics, which is strongly nonlinear and interconnected, with a high number of structural constraints (e.g., strings must always be in tension, strings and bars should not collide with one another). The second problem is the complexity of actuation and sensing. Indeed, tens of actuated strings would be present in a tensegrity manipulator, and the position and velocity of at least a significant fraction of the bars should be measured or estimated. Our project aims at defining a general paradigm for the motion planning, state estimation, and closed-loop control of tensegrity robot manipulators based on numerical optimization and machine learning.

This research activity is funded by Nazarbayev University faculty development research project “Motion Planning and Control of Tensegrity Robots” (2020-2022), and is in collaboration with the NU ARMS Lab.