Energy-based local forward and inverse kinematics methods for tensegrity robots

Abstract

This paper describes an energy-based forward kinematics method for tensegrity robotics structures. The method can be implemented as a nonlinear optimization procedure and solved with standard optimization algorithms. Based on this result, the paper proposes a local linearization-based approach to inverse kinematics for tensegrity robots, for which a gradient descent-like algorithms is formulated. A preliminary study of the algorithm numerical behavior is given, and examples of the use cases for both forward and inverse kinematics methods are provided. It is shown that the method allows planing trajectories in the Cartesian space for the end effector and discover the desired law of change of the rest lengths of the actuated elastic elements.