Indoor Position Estimation System for Wide Area Augmented Reality Applications

An important problem in wide area augmented reality applications is to localize the user accurately so as to superimpose metadata in the user’s field of view. This task is easier in outdoor environments since GPS is available. Also, for small area AR applications consisting of a scene with a great deal of texture, computer vision techniques can be used to determine a user’s pose. In this paper, we propose an end-to-end system which can be used to (a) construct 3D models and associated 2D floor plan of large scale indoor environments in a fast, automated, scalable way; (b) construct multiple sensor e.g. WiFi and imagery signature databases for the same environment; (c) use the reference databases in part (b) to recover position of users wearing an AR glass or carrying a mobile device equipped with a camera and WiFi sensor. The system consists of a man portable backpack of sensors carried by an operator inside buildings walking at normal speeds. The sensor suite consists of laser scanners, cameras and an IMU. Particle filtering algorithms are used to recover 2D and 3D path of the operator, a 3D point cloud, the 2D floor plan, and 3D models of the environment. The same walkthrough that results in 2D maps also results in multi-modal sensor databases, in our case WiFi and imagery. The density of the resulting WiFi database in our system is considerably higher than existing systems due to continuous, rather than stopand-go, WiFi signature acquisition. We use particle filtering algorithms in an Android application to combine inertial sensors on the mobile device or glass, with 2D maps and WiFi and image sensor databases to localize the user. Experimental results for the second floor of the electrical engineering building at UC Berkeley campus will be shown.