Proposal of Indoor AR Navigation System Using SLAM for Location Search

In recent years, the increased computing power of mobile terminals has enabled users to recognize images accurately and in detail. In particular, augmented reality (AR) spaces created from camera shots by mobile terminals are being applied to navigation technology that guides users to their destinations indoors. Unlike outdoor systems, most indoor AR navigation systems cannot use the Global Positioning System (GPS). Therefore, AR navigation systems based on simultaneous localization and mapping (SLAM) have been proposed. SLAM uses physical information such as AR markers and distinctive shapes to simultaneously estimate the self-position of a terminal used indoors and create an environment map, which gives map information of the area around a movement path. However, this system requires the proper placement of a large number of AR markers. In addition, as the travel distance increases, errors occur in the self-position estimation. In this paper, we propose an indoor AR navigation system that uses multiple environment maps based on both SLAM and AR markers. The proposed system recognizes areas where navigation is possible by applying the environment maps created using SLAM, and then it presents a route in AR that excludes object areas that would present obstacles when moving indoors. Accordingly, users can move to their destinations while avoiding such obstacles. We designed and implemented a prototype of the proposed system and evaluated its performance. From the evaluation results, we confirmed that the proposed system can identify multiple rooms using AR markers by performing self-position estimation based on the feature points given in the environment map.