Abstracting routes to their route-defining locations. (January 2022)

• Record Type: Journal Article
• Authors: Teimouri, Fateme; Richter, Kai-Florian
• Is Part Of: Computers, environment and urban systems. Volume 91(2022)
• Publication Date: 2022-01
• Subjects: Wayfinding -- Route generalization -- Prominence -- Agent-based modeling -- Spatial learning; City planning -- Data processing -- Periodicals; Regional planning -- Data processing -- Periodicals; 303.4834
• DOI: 10.1016/j.compenvurbsys.2021.101732
• Languages: English
• ISSNs: 0198-9715

Abstract

Abstract : Highlights: Highlighting prominent locations (Landmarks, Streets, Major turns) along the route. Abstracting the route to its characteristics with different levels of abstractions. Beneficial to generate route instructions that are tailored toward the wayfinder's knowledge. Abstract: Today's navigation assistance systems provide turn-by-turn instructions, which only focus on the next decision to take without offering any larger context. Information presentation is uniform and disconnected, i.e., any local instruction is equally important and not linked to any other decisions or the overall route context. This differs from how people usually give instructions and hinders spatial learning. In order to (re-)establish this larger context, we present an approach to identifying those locations along a route that define its characteristics, termed route-defining locations . These are prominent, easily recognized locations, which help relating the route to an environment's overall structure and a navigator's existing knowledge about the environment. The approach allows for determining route-defining locations on different levels of detail. Thus, at the same time it offers a mechanism for simplifying (or abstracting) a route. In this paper, we particularly focus on the latter aspect, presenting in detail the approach for identifying route-defining locations. For a given route, we, first, simplify its shape to extract those turns along the route that characterize its overall shape. Then, we rank landmarks and streets along the route based on their prominence. Finally, we include in the simplified route those turns that correspond to the locations of the most prominent landmarks and streets. The result is a set of route-defining locations extracted from the route shape, and prominent landmarks and streets along the route. In an agent-based simulation, we then evaluate the approach's ability to abstract a route to its characteristics, i.e., the defining locations. Results show that, indeed, our approach is effective in that respect, but success depends on 'matching' abstraction to an agent's knowledge about the environment.