Robotic Assembly Planning of Tensegrity Structures

Abstract

Tensegrity structures provide durability and stability with minimal weight. Thanks to these properties, use of tensegrity structures for space applications is becoming an active research area. One of the main challenges is the automated construction of such structures in space. In this thesis, we present a framework for the automated assembly planning of tensegrity structures using an industrial manipulator (Staubli TX90XL) equipped with a purpose-specific end-effector. We leverage the recent advances in sampling-based motion planning to create the motion plans (i.e. the reference trajectories for the manipulator joints). Specifically, we divided the assembly planning problem into three stages. In the first stage, the initial position and orientation of the given tensegrity structure with respect to the assembly robot is determined using a motion-planning integrated forward elimination search. In the second stage, a feasible assembly sequence of the strings is found using backward disassembly search. Lastly, individual robot motion plans for attaching each string to the bars for the given tensegrity configuration is generated using RRT* algorithm. The efficacy of the framework was demonstrated using an extensive set of simulation and real-world experiments dealing with the assembly of a 3-strut and 9-string tensegrity prism structure, which can be utilized as a building block of complex tensegrity structures.