Empirically measuring salience of objects for use in pedestrian navigation

Abstract

Auch wenn sie es vehement von sich weisen würden, so haben meine Eltern großen Anteil an dieser, meiner Dissertation. Sie gaben mir von Kindesbeinen an die Freiheit, mich mit dem Nachdenken auseinanderzusetzen. Glaubt mir, Mama und Papa, ohne diese Freiheit wäre ich heute nicht der, der ich bin.Schließlich danke ich meiner über alles geliebten Frau Anita. Wir sind fast genau so lange verheiratet, wie es gedauert hat, bis ich nun mein Promotionsverfahren erfolgreich abschließen konnte. In diesen Jahren haben wir einiges zusammen erlebt und durchgemacht.Du, Anita, hast immer zu mir gehalten und mich in Allem und mit Allem unterstützt.Ohne Dich, Anita, wäre diese Arbeit undenkbar.Dir, Anita, widme ich diese Arbeit.vi A B S T R A C T background Humans usually refer to landmarks when they give route directions to pedestrians. One of the reasons why current mobile pedestrian navigation systems do not yet mimic this mode of communication is the lack of available data sources. The usefulness of a crowd-sourced data acquisition approach to overcome this problem has long been mooted. However, to date no empirically sound way of measuring the salience of objects by means of surveys exists.goal Given this background, this doctoral work has three goals:1. To achieve a sound way of measuring salience and its subdimensions, i.e. visibility in advance, cognitive salience, prototypicality, structural salience, and visual salience based on taking dimensions revealed in earlier studies systematically and simultaneously into account.2. To find subgroups of visual features among the large number of visual attributes known from the literature.3. To find the most important subdimensions of salience by means of estimating two different structural equation models. Model I is based on assumptions of independence among subdimensions, whereas model II reflects hypotheses of mediation.Taken as a whole, achieving these goals will foster both, the advancement of theories of salience and landmark acquisition methods.methodology A large scale, in-situ experiment was implemented, trying to overcome weaknesses of earlier attempts made to estimate salience. An vii appropriate sample size of buildings and non-buildings was calculated a priori (n obj = 360). Objects were randomly selected based on their geographical coordinates and randomly grouped into n r = 55 routes. Participants were required to rate objects by means of a survey. The questions were derived from empirical evidence found in earlier studies. Each route was walked by two different participants (n = 112), id est (i.e.) two ratings per object were collected for data analysis.findings Model I and model II were analyzed using PLS Path Modeling and consistent PLS Path Modeling, respectively. The measurement models proposed showed a good fit, although some weaknesses were identified for prototypicality and cognitive salience. Geometrical aspects as well as features like (visual) age turned out to have a stronger impact on visual salience than color. Model I did not yield reasonable struct...

Task-Oriented Evaluation of Indoor Positioning Systems

“…salient landmarks -are located. For determining salient landmarks along the route, we followed the approach described in [10]: 19 persons rated 32 objects in the test area regarding different aspects of their salience. We selected the objects with the highest predicted overall salience as landmarks for the navigation instructions in our experiment.…”

On the Right Track: Comfort and Confusion in Indoor Environments

“…Because landmark-based wayfinding is known to be effective, landmarks should be included in route instructions [55,56]. Current studies focus mainly on the saliency of landmarks (e.g., [56][57][58][59]).…”

Navigation in Indoor Environments: Does the Type of Visual Learning Stimulus Matter?

“…Some scientists believe that people are better able to remember visual elements that closely depict reality [31][32][33]. Higher levels of realism in visualizations can also mean higher ecological validity if we are modeling a real environment [34,35]. Other authors believe that higher levels of realism increase cognitive load and exhaust memory capacity more quickly [36][37][38][39].…”