Abstract
In an unfamiliar environment we spot and explore all available information which might guide us to a desired location. This
largely unconscious processing is done by our trained sensory and cognitive systems. These recognise and memorise sets of
landmarks which allow us to create a mental map of the environment, and this map enables us to navigate by exploiting
very few but the most important landmarks stored in our memory. In this paper we present a route planning, localisation and
navigation system which works in real time. It integrates a geographic information system of a building with visual
landmarks for localising the user and for validating the navigation route. Although designed for visually impaired persons,
the system can also be employed to assist or transport persons with reduced mobility in way finding in a complex building.
The developed navigation framework offers robust localisation, even in cases where it is difficult to
differentiate regions on the basis of reference objects that are present. The navigation history enables autolocalisation,
by selecting the region that combines the largest number of nearby objects in the history.
However, the system supported in the estimated direction information to the next move (vector for
displacement). Localisation usually fails when the user deviates from the assumed or expected path. The correct
location is then restored as soon as sufficient navigation history (objects) or a single significant object exists.
False positive and negative detections may also interfere negatively. In the future a verification technique that
can rule out false detections by analysing the proximity relations with the other navigation history objects.
Nevertheless, most planned routes could be followed from the starting to the destination, even when these were
on different floors.
On-going work concerns including more landmarks in the database such that all spaces can be covered more
densely, and detecting doors and stairs from oblique viewing angles, such that the user can limit the pointing
angles of the camera to ±45º from the front. The final goal is a system which only employs a SmartPhone with
a built-in camera, worn by a strap around the neck.