Route Adaptive Selection of Salient Features

Abstract

AbstractHuman communication on wayfinding makes extensive use of landmarks. With a formal model of salience, route planning services can include landmarks as well. Such a model was presented considering visual, semantic, and structural properties of spatial features. This model measures saliency independent from a given route. Our hypothesis is that an additional factor is cognitively relevant for the selection of appropriate salient features: advance visibility for a person approaching a destination point. We will propose a computational measure for advance visibility. The new measure is used to identify suited salient features at route decision points: a feature is suited for a wayfinding instruction if it is (a) salient, and (b) in advance visible. The relevance of advance visibility is tested by a comparison of wayfinding success with instructions made with and without this additional measure. Computational effort is observed to check feasibility.KeywordsSalient FeatureStreet SegmentVisibility AreaDecision PointVisibility GraphThese keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.