Learning of mobile working for humanoid robots by observation and communication

Abstract

Actual environments where people live and do their tasks are not always arranged for robots. When the robots have to execute their task, for example guiding a person to a destination, it uses a database like map, but as the task environment may often change, the database should be maintained simultaneously. This will be a heavy burden for the owner of the robot, so, some autonomous methods like observation learning are required. In this report, we propose an observation and communication method so that a humanoid robot can create and maintain a map for navigation.