GoIn - An Accurate 3D InDoor Navigation Framework for Mobile Devices

Performing a building-level positioning using WLAM and cellular information is a well-known methodology which was suggested and implemented by many researches. In this paper we present a general framework for accurate indoor positioning and navigation which improves the expected accuracy to a sub-meter error rate. The main algorithm is based on a modified particle filter which combines RF finger-printing, odometry, visual landmarks and map constrains. The accuracy improvement is achieved by using a low resolution camera to track dominant landmarks such as lights. The use of 'glowing-markers' allows one to accurately map relatively complex indoor buildings with a compact representation. The suggested method [BmS15] was implemented and tested on android based mobile devices. Our tests indicate a robust sub-meter 3D positioning at 10 - 30Hz with a fairly low energy consumption.