Dynamic Modeling and Controlling for the Crawling Motion of the 6-strut Tensegrity Robot

Abstract

The 6-strut tensegrity robot (TR-6) is a novel deformable mobile robot, and it can generate different gaits by self-deformation to adapt to various terrains. As the shape of the TR-6 is similar to a sphere, unexpected falling often occur when the robot is climbing a slope with rolling gait. By self-deformation, the TR-6 can generate crawling gait and move forward steady. As the topology of the TR-6 is complicated, the components of the structure are highly coupled, and the TR-6 is a multi-input system, there is no existed method for solving the control problem of the crawling gait for TR-6. This paper presents a control method for the crawling gait of TR-6 based on the genetic algorithm (GA). Firstly, a dynamic model of the TR-6 is built by the Newton-Euler method. Secondly, based on the TR-6 dynamic model, an optimal initial configuration and the range of the control parameters for avoiding the unexpected rolling during crawling motion are determined. Lastly, by converting the multi-input control problem to an optimization problem and solving the optimization problem by GA, a group of control parameters are chosen to control the TR-6 crawl forward steady and fast. The control method is realized by C++ code, and a group of control parameters are given by the numerical simulation. The experiments verifies the result of simulation.