Abstract

The seminar will present two areas of investigation that have led to the development of features for future Navigation systems. We have concentrated on the application to road vehicle users, but some of the techniques may be more generally useful.

The first investigation looked at what might be possible if all navigation units kept a complete log of their movements. This led to the development of route planners that take into account the familiarity of the user with particular route segments, navigation aids that operate without user input by predicting future behaviour from past traces, and desktop applications that can use the vehicle history to correlate many types of object by location.

The second investigation looked at finding a “good set of routes” from one place to another. The set is hard to define mathematically, but we informally refer to this when we say “you could go this way, or that way, or even via there”. Can we find an algorithm that will generate the good alternative routes that an expert navigator might consider, and how would we present these to the user?