Identifying motion and interest patterns of shoppers for developing personalised wayfinding tools

Abstract

The development of wayfinding and information tools for pedestrians faces several challenges. In contrast to common navigation tools used for vehicles, navigation services for pedestrians must fulfil more complex requirements in order to be accepted. For pedestrians, the shortest path does not always represent the optimal route for an individual's purposes, as studies have revealed that people often forgo to take the shortest path and prefer the 'most beautiful', 'most convenient' or 'safest' path. People exploring a new environment on foot would therefore especially benefit from systems providing information concerning route qualities, interesting facilities in the vicinity and other useful location-related suggestions. As part of the scientific project 'UCPNavi', we investigate group-specific spatio-temporal behaviour and related influence factors of shoppers in order to identify homogeneous behaviour types based on motion patterns and information requirements. We use an eclectic approach combining several complementary empirical methods of data collection and analysis to thoroughly comprehend pedestrian spatio-temporal behaviour. In this contribution, we present results of motion and interview data analysis based on data collected from more than 100 participants during two empirical phases on a shopping street in Vienna. We further introduce an initial pedestrian typology comprising five different homogeneous behaviour types based on qualitative-interpretative and quantitative-statistical data. Types are described according to characteristic attributes identified by route choice behaviour, walking patterns and interest foci. The relevant factors include velocities, stopping behaviour, categories of visited facilities and individual preferences. The resulting typology of lifestyle-based pedestrian mobility styles and the identified characteristic attributes can serve as a basis to create pedestrian interest profiles in ubiquitous environments and to customise navigational and environmental information for mobile applications in order to fulfil individual needs.

Classifying users of mobile pedestrian navigation tools

“…Using spatio-temporal data of pedestrian shoppers, pedestrians have been classified as utilitarian, hedonistic, convenient, swift, and discerning [5]. Taking tool usage into account and using eye tracking with verbal data, usage of a map-based tool yielded three types of pedestrian users: Constant Support and Information, Independent and Attentive, and Least Effort and Inattentive [9].…”

“…While car drivers can generally be relied upon to follow turn-byturn directions of an in-car GPS system, pedestrians enjoy greater degrees of freedom and have far more complex requirements to fulfill [5]. Consequently, there are a wide variety of navigation tools continually being introduced for people on foot, from traditional maps with dynamic self-location features, to Augmented Reality (AR) tools.…”

Spatio-temporal facility utilization analysis from exhaustive WiFi monitoring

“…Moreover, the already available wireless infrastructures in large building complexes, like hospitals, enable the collection of large data sets of WiFi measurements that can be used not only to analyze the network's performance and usage, as proposed in earlier work among others [8][9][10], but potentially also the density and flow of people within the building. Compared to earlier approaches based on Bluetooth, in urban [11] or indoor settings [12], or based on video in indoor settings [13], the use of WiFi comes with lower setup costs, due to the existing deployment, for monitoring complete large-scale building complexes. In our related work [14,15], we propose WiFi-based methods for detecting commonly used routes and estimating travel times for specific persons based on WiFi position traces collected for these persons.…”

Classifying pedestrian movement behaviour from GPS trajectories using visualization and clustering

“…When this knowledge is related to human movement, it can be useful in location based services (Millonig and Gartner, 2011), emergency management (Zheng et al, 2009), traffic management (Castro et al, 2012), health (Landau et al, 2009) and urban planning (Van der Spek, 2010). In urban environments, we can gain particularly valuable knowledge from data on pedestrian movement.…”