Self-Position detection system using a visual-sensor for mobile robots

摘要

We have been developing an intelligent mobile robot for use as an office building secretary or aid during the day and a security guard or maintenance engineer, e.g., for cleaning floors, at night. Since the robot works and moves autonomously among people in an office environment, it must be able to recognize its own location and environment. We proposed two types of self-position detection based on a visual sensor. One is global self-positioning (GSP) by recognizing a room number. The other islocal self-positioning (LSP) calculating the relationship between the robot and three light landmarks such as two exit lamps and a fire hydrant lamp in corridors. Experiments verified the effectiveness of the robot's self-position detection.