Sensorimotor Information flow in Genetic Regulatory Network driven control systems

摘要

We present results from applying information theory based measures to simulated robots controlled by evolved Genetic Regulatory Networks. Measuring the information flow across sensory and effector surfaces, we create an information profile illustrated using area-proportional Venn diagrams. We examine the relationship between this information profile and various elements of the overall system including the GRN controlling the robot, the environment, the nature of the task for which the GRN controller has been evolved, and evolutionary fitness.