Over the past several decades, there has been a considerable amount of research conducted on the nature and quality of route directions, including those in mixed environments (Fontaine & Denis, 1999; Lovelace et al., 1999). We extend this work to situation

Knowledge Based Schematization of Route Directions

Samvith Srinivas1, Stephen C. Hirtle2

1University of Pittsburgh

hirtle@pitt.edu

University of Pittsburgh

sas29@pitt.edu 2

Over the past several decades, there has been a considerable amount of re-search conducted on the nature and quality of route directions, including those in mixed environments (Fontaine & Denis, 1999; Lovelace et al., 1999). We extend this work to situations where wayfinding tasks take place in an environment that is composed of a familiar part and an unfamiliar part, which we refer to as a partially familiar route. We introduce a formalization that models routes of this nature, conduct an em-pirical study to investigate whether partially schematized directions are preferred over complete directions for such routes, and conclude with observations about the usefulness of the formal approach for future implementation in a wayfinding system.

CONCEPTS FOR CODING FAMILIARITY

We use a graph theoretic approach to model the heterogenity of route knowledge. The terminology we use to represent routes is derived from work on wayfinding choremes (Klippel, 2003a, 2003b; Klippel et al., 2005).

To Klippel’s original framework, we introduce the concept of a partially familiar route, or knowledge route (<kroute>). A partially familiar route is one which incorporates a familiar region (<K>) and an unfamiliar region (<N>) along the same route (Figure 1). This gives the most basic form of a partially familiar route. The braces indicate that the order of <K> and <N> can be interchanged.

<kroute> ::= <O> {<K> <N>} <D>