How does environment bias behavior of autonomous robots with evolutional perception

In this paper, concerning with the emergence of intelligent behaviors of some artificial systems like an autonomous robot, we put an emphasis on the importance of their continuous interactions with an environment, and consider about how an environment affects on their forming behavioral rules that are needed for their survival. For these purposes, we at first develop an evolutional algorithm for an autonomous mobile robot's behavior formation, and discuss about how the ways of interactions bias the robot's learning through a number of experiments. Based on these, we investigate into possible instruction strategies for an variety of environments as well as into appropriate learning strategies of maintaining a diversity for a robot.