Reaching hidden objects based on memory of environmental states and robot's movement and manipulation

Abstract

We aim at a robot that searches domestic objects with various manipulations of environments such as opening a drawer. A robot memorizes places of them and reaches them again even if they are hidden, e.g. an object in a drawer. We think that the location of the robot should be represented not based on Cartesian coordinates but based on the state of each environment, because robot's action to reach hidden objects varies by the state of each environment. The robot will reach the place of the memorized object by the action based on the integrated memory of movement and manipulation when the robot was searching the object. This paper describes the concept and the basic method of memorizing robot's movement and manipulation integratedly, and generating the action to reach hidden objects based on the memory.