Visual environment mapping for mobile robot navigation

Abstract

In order to successfully navigate mobile robots in an unknown environment, an appropriate navigation strategy must be employed. Such a strategy certainly relies on an effective navigation map based on which robot trajectories can be well planned. The navigation map can be built based on a sparse map constructed by a visual simultaneous localization and mapping (VSLAM) algorithm. This paper presents a map conversion strategy to build a 3D navigation map employing a scout robot equipped with an onboard camera. According to the 3D navigation map, ground path planning can be accomplished by a proper navigation strategy. The proposed map conversion strategy has been successfully validated by experimenting with a mobile robot in an indoor laboratory environment.