Evolving Sensor Morphology on a Legged Robot in Niche Environments

Abstract

This paper discusses the issue of evolution of morphology and automatic design, specifically evolving sensor morphology on a legged hexapod robot in niche environments. The evolution of sensor morphology in different environments, in particular, type of sensor, angle of heading and its effect on controller complexity for a simulated hexapod robot are described. This automatic design method enables the system to decipher relevant stimuli in an environment, increases the efficiency of the robot and also indirectly alters the controller of the robot to take advantage of the characteristics of a given environment.