Towards An Indoor Navigation System Using Monocular Visual SLAM

This paper presents a novel implementation of an augmented reality-aided indoor navigation system for mobile devices. The proposed system uses the device's camera to scan the environment, generate an abstracted 3D map of the indoor environment, and transfer it to a remote server. The 3D map of the indoor environment is achieved through a tracking and mapping ARKit module. Once the indoor map is stored in the server, it can be accessed simultaneously by multiple devices for localization and navigation. Leveraging Unity assets and the directions retrieved from the server, the application computes the shortest distance between the source and destination, and displays AR-based direction markers for navigation assistance.