Hybrid map building for personal indoor navigation systems

Tracking the positions of people in large indoor spaces is important, since it enables a range of applications related to security, indoor navigation and guidance. This paper proposes a personal indoor navigation system based on hybrid map, containing geometric as well as symbolic information. In this way the same map can be exploited to guide and localise the user efficiently during navigation. The hybrid map is built using floor plans of the environment. It is a topological graph capturing the connectivity of complex indoor environment and it is retrieved by applying image-processing techniques. Some additional metric information are added to make the map suitable for quantitative localisation. Semantic features are considered to improve user readability