AI for Innovation Design of Tensegrity Mobile Robot

Tensegrity mobile robots have advantages of high stiffness-to-mass ratio and superior structural compliance, making them a hot research topic recently. However, it is still challenging in innovation design of the tensegrity mobile robots under specific terrains. In this work, an AI-based innovation design method for the tensegrity robots is proposed. Firstly, starting from configurations of tensegrity unit, a library for movable tensegrity units is established combining permutation and force density method. This can provide foundations for the AI-based generation of tensegrity robots. On this basis, genetic algorithm, which includes the selection, crossover and mutation of genotype, is introduced to generate the configurations of tensegrity robots. Finally, by continual selection and evolution under the specific terrains, the configurations of the robots that can successfully pass through the terrain and have high fitness will be obtained. This work provides a useful reference for the application of artificial intelligence (AI) in the field of innovation design of tensegrity mobile robots.