Content area
Full text
1.
INTRODUCTION
Navigation services have become one of the most widely used types of Location-Based Services (LBS). Many of the currently available navigation systems provide users with turn-by-turn navigational instruction, in which directions are associated with distances, e.g. "turn right after 250 metres". Such a navigational strategy works very well for machine/robot navigation, as sensors sense and measure distances and headings. However, people give navigational instructions in a slightly different format. People associate directions with visual cues, such as "turn left at the church". Such visual cues utilise easy-to-recognise and unique features and objects.
This strategy might be more attuned to the interests of pedestrians, since they move with relatively lower speed; therefore they can notice visual landmarks easily. In addition, such unique and easy-to-recognise features help people to memorise the path they have taken and also give a better understanding of their surroundings, especially when they are exploring unfamiliar environments. Landmarks are also interesting features in their own right for many people, such as tourists.
Landmarks are available both inside and outside buildings. Therefore any positioning technique that localises users with respect to landmarks, whose locations are usually fixed, can potentially provide seamless (indoors and outdoors) positioning solutions. Seamless indoor and outdoor positioning is one of the most challenging parts of pedestrian navigation. Seamless navigation is the topic of many research projects (Cheng et al., 2014; Zhang et al., 2012; Li et al., 2013; Hansen et al., 2009). In this regard, landmarks can be used as reference points in positioning techniques, which may then provide a seamless positioning solution. This paper proposes and implements one technique, which uses photographs of landmarks to localise users.
In contrast with drivers, pedestrians have a higher degree of freedom in their movements. They can walk across open areas such as squares, parks, grasslands or pedestrian malls, which can be traversed freely in any direction (Gaisbauer and Frank, 2008). As current turn-by-turn navigational instructions designed to be given to vehicle drivers are mostly based on graph-based or street network-based algorithms, this way of navigating is not fully suitable for pedestrians as they do not only move on streets (Pielot and Boll, 2010; Amirian et al., 2015). Therefore landmarks can be used to provide seamless positioning and more desirable navigation...