An acquisition of a narrow path going through behavior for a distributed autonomous swimming robot

To achieve a “strong” autonomous mobile robot which can acquire adaptive behavior through learning in the complex real world, constructing the robot as a distributed autonomous system can be one solution. It is thought that such a system has many advantages over other systems in terms of adaptability to the environment and so on. However, the potential of this type of system has yet to be demonstrated in experiments under real-world conditions. So, we developed a distributed autonomous swimming robot which had redundant linkage structure, and in which each link were independently controlled. The reason why we chose a water environment as the robot´s work space is that the environment is uncertain and unstable in the real world. Under these settings, we conducted an experiment to determine whether the robot could complete target approaching including obstacle avoidance which is a necessary function for autonomous mobile robots. In this paper, the obstacle is a narrow path so the robot can´t go through it by chance. A behavior-acquiring experiment for target approaching was implemented, and after that, the capability of narrow path following was tested by using acquired action patterns. As a result, from a fairly simple coding, the robot acquired reproducible target-approaching behavior using only local learning, and completed narrow path following behavior in a conditional environment