Environmental Adaptation of Robot Morphology and Control through Real-world Evolution

Robots operating in the real world will experience a range of different\nenvironments and tasks. It is essential for the robot to have the ability to\nadapt to its surroundings to work efficiently in changing conditions.\nEvolutionary robotics aims to solve this by optimizing both the control and\nbody (morphology) of a robot, allowing adaptation to internal, as well as\nexternal factors. Most work in this field has been done in physics simulators,\nwhich are relatively simple and not able to replicate the richness of\ninteractions found in the real world. Solutions that rely on the complex\ninterplay between control, body, and environment are therefore rarely found. In\nthis paper, we rely solely on real-world evaluations and apply evolutionary\nsearch to yield combinations of morphology and control for our mechanically\nself-reconfiguring quadruped robot. We evolve solutions on two distinct\nphysical surfaces and analyze the results in terms of both control and\nmorphology. We then transition to two previously unseen surfaces to demonstrate\nthe generality of our method. We find that the evolutionary search finds\nhigh-performing and diverse morphology-controller configurations by adapting\nboth control and body to the different properties of the physical environments.\nWe additionally find that morphology and control vary with statistical\nsignificance between the environments. Moreover, we observe that our method\nallows for morphology and control parameters to transfer to previously-unseen\nterrains, demonstrating the generality of our approach.\n