Research on Motion Evolution of Soft Robot Based on VoxCAD

The shape changes of soft organisms demonstrate the survival rules in the evolution of self life. It is important for the transform control of soft robots about how to envolve suitable for the shape demands. In this paper, the compositional pattern producing networks (CPPN) algorithm was used to evolve soft robot. By taking simple random functions as genotype inputs, the functions can be weighted combinations to generate the desired phenotype, which mapping relationship between genotypes and phenotypes can be achieved. The VoxCAD simulation software was used to build the three-dimensional topological structure of soft robot and the evolution process of the virtual life was realized by using the specific rule shape to simulate the real environment. The evolutionary analysis of the four-legged walking soft robot was carried out in the simulation experiment, which the effectiveness of the method was verified.