Landmark weights - an alternative to spatial distances in shortest route algorithms

Abstract

ABSTRACT Although numerous studies have shown that landmarks are important for navigation, almost all navigation systems implement a shortest-route algorithm without considering landmarks. Which options do we have to integrate landmarks into a route calculation? How would the resulting routes differ from shortest routes? We propose a weighting method for Dijkstra’s shortest route algorithm to generate “landmark routes” and compare them to their corresponding shortest routes. We show that the extra distance and time needed to walk the landmark routes is acceptable in most of the routes. The main contribution is a thorough discussion of the differences between the two types of routes. Since the results are promising, we discuss variations in computing the weights as well as recommend human subject tests.