Floor-types identification method for wheel robot using impedance variation

Abstract

Wheel robots have been widely used in indoor. An indoor room consists of various floor-types such as wood, carpet, marble and so on. The various floor-types cause odometry error and energy loss due to slip. Floor-type identification which allows a robot to detect changing floor-types and to adapt control and motion planning has been needed to solve this problem. In this paper, we propose sensing method for wheel robot to identify floor-types by using impedance variation. The impedance variation measures a voltage variation of a side brush motor depending on floor-types. The original side brush function is to sweep the dust in the vacuum cleaning robot. This approach does not require any sensors and estimates floor-types from the impedance variation. We present and analyze experimental data collected on three different floor-types: wood, marble and carpet. As a result, we verify the fact that the floor-types can be identified by the proposed method.