다중 센서를 이용한 자율주행로봇의 위치보정 및 장애물 회피

초록

This paper presents algorithms for localization and obstacle avoidance of an autonomous mobile robot using a monocular vision system and an ultrasonic sensor to implement a service robot with map information in corridor environments. This paper implements a vision system modeling and calibration to perceive corridor environments. As results, it is allowed to obtain real world coordinates from an image of vision system, and it is possible to avoid obstacles and localize the robot position based on extracted features from the image. This paper also presents an image partition method and a limited image processing method to reduce processing time. As a preliminary experiment, the system acquires corridor information with a monocular vision sustem, and the result demonstrates the stability of the proposed methods for a service robot.