수중 초음파 거리 센서를 이용한 수중 로봇의 2차원 지도 확장 실험

Abstract

This study focused on autonomous exploration performed by an underwater robot using an acoustic range sonar. A basic goal in autonomous exploration is map expansion, which involves mapping large areas without loss in region. Map expansion conflicts with the localization of the underwater robot. As the created map is expanded, the accuracy of the robot position may be decreased, and this leads to reduction in the map accuracy. Thus, the technique of localization of the underwater robot is also crucial in autonomous robot exploration. Herein, we propose a method for underwater robot exploration wherein the robot conducts a tradeoff between the map expansion ratio and the position accuracy, selects the one with higher priority, and decides upon the subsequent strategy. An acoustic range sonar is adapted to create the occupancy grid map. The areas in the created map are designated as free, occupied, or unknown regions based on the probability of occupancy. This information is then used to determine a path to the frontier, which is the new search point between the free and unknown regions. Robot localization is performed simultaneously during the exploration of a search area and map creation by a robot. If the accuracy of the robot position is less than an assigned threshold, the robot revisits the artificial landmark to improve its position accuracy using landmark recognition by the imaging sonar. Real-time experiments w