In smartphone-based pedestrian navigation systems, detailed knowledge about user activity and device placement is a key information. Landmarks such as staircases or elevators can help the system in determining the user position when located inside buildings, and navigation instructions can be adapted to the current context in order to provide more meaningful assistance. Typically, most human activity recognition (HAR) approaches distinguish between general activities such as walking, standing or sitting. In this work, we investigate more specific activities that are tailored towards the use-case of pedestrian navigation, including different kinds of stationary and locomotion behavior. We first collect a dataset of 28 combinations of device placements and activities, in total consisting of over 6 h of data from three sensors. We then use LSTM-based machine learning (ML) methods to successfully train hierarchical classifiers that can distinguish between these placements and activities. Test results show that the accuracy of device placement classification (97.2%) is on par with a state-of-the-art benchmark in this dataset while being less resource-intensive on mobile devices. Activity recognition performance highly depends on the classification task and ranges from 62.6% to 98.7%, once again performing close to the benchmark. Finally, we demonstrate in a case study how to apply the hierarchical classifiers to experimental and naturalistic datasets in order to analyze activity patterns during the course of a typical navigation session and to investigate the correlation between user activity and device placement, thereby gaining insights into real-world navigation behavior.

Abstract. Knowledge about real-life user behaviour is an important factor for the design of navigation systems. Prompted by the observation that users tend to use our campus navigation app in unexpected ways, we present a naturalistic study of navigation logs. The data set consists of sensor and interaction data from over 4600 sessions, collected over a span of several months from hundreds of users. In our analysis, we demonstrate how the core concepts from navigation literature, i. e. wayfinding and locomotion, can be observed, but also point out differences to previous studies and assumptions. One of our main findings is that the application is mostly used to plan routes in advance, and not to navigate along them. Furthermore, detailed case-studies of actual navigation sessions provide a unique insight into user behaviour and show that persons are often not focused on their navigation task but engaged otherwise. Based on these results, we formulate design implications that do not only apply to future iterations of our application, but can be seen as best practices for pedestrian navigation apps in general.

Many smartphone-based indoor positioning systems rely on pedestrian dead reckoning for fine-grained position tracking. In this paper, we show how one of its major shortcomings - the accumulation of errors over time - can be effectively overcome in a navigation setting by detecting door transitions along the route. Using only the sensors included in a handheld smartphone and different on-device machine learning techniques, door transitions can be classified correctly in up to 86% of all cases. The system runs in real-time on current smartphones and - when integrated into our baseline particle filter - improves overall positioning performance significantly, as the subsequent evaluation on a realistic navigation data set shows. A detailed analysis of several edge cases illustrates the concept and provides insights into the remaining challenges.

Pedestrian Dead Reckoning (PDR) plays an important role in many (hybrid) indoor positioning systems since it enables frequent, granular position updates. However, the accumulation of errors creates a need for external error correction. In this work, we explore the limits of PDR under realistic conditions using our graph-based system as an example. For this purpose, we collect sensor data while the user performs an actual navigation task using a navigation application on a smartphone. To assess the localisation performance, we introduce a task-oriented metric based on the idea of landmark navigation: instead of specifying the error metrically, we measure the ability to determine the correct segment of an indoor route, which in turn enables the navigation system to give correct instructions. We conduct offline simulations with the collected data in order to identify situations where position tracking fails and explore different options how to mitigate the issues, for example through detection of special features along the user’s path or through additional sensors. Our results show that the magnetic compass is often unreliable under realistic conditions and that resetting the position at strategically chosen decision points significantly improves positioning accuracy.

The performance of indoor positioning systems is usually measured by their accuracy in meters. This facilitates the comparison of different systems, but does not necessarily give information about how well they perform in real-life scenarios, e. g. during indoor navigation of walking persons. In this paper, we present a task-oriented evaluation that adapts the idea of landmark navigation: Instead of specifying the error metrically, system performance is measured by the ability to determine the correct segment of an indoor route, which in turn enables the navigation system to give correct instructions. We introduce the area match metric in order to identify areas where positioning proves problematic. In order to evaluate the described metric, we use a pedestrian dead reckoning approach to compute indoor positions. Without any external correction, the correct segment of the test route is identified in 88.4% of all trials. Based on these results, we explore options how to identify and predict erroneous situations during the navigation process as well as beforehand.

Any unassisted decision a user has to make can be difficult, even if it entails the simple task of selecting which movie to watch. The sheer volume of movies offered by streaming platforms makes this task all the more difficult and time consuming. Many platforms attempt to combat this problem through recommendation systems. These however seem more likely to be making wild suggestions than being a constructive aid to the selection process. In order to offer more accurate recommendations, we propose a system that is based on a user's current emotion, which is matched with the sentiments contained in the movies' spoken language. A study involving our newly designed mobile sentiment-based movie recommender named 'FROY' shows highly promising results. As it turns out, sentiment analysis of spoken language leads to appropriate recommendations.

In dieser Arbeit werden Theorien und Erkenntnisse über das

Fusgangernavigationssysteme benutzen fur die Zielfuhrung idealerweise auffallige Objekte (Landmarken). Die Nachteile bisheriger Verfahren zur Gewinnung dieser Objekte konnen durch die Erfassung in alltaglichen Navigationssituationen mithilfe einer Smartphone-Applikation umgangen werden. Im Folgenden werden notige Vorarbeiten zur Umsetzung einer solchen Anwendung beschrieben: Zunachst mussen verschiedene Datenbestande auf ihre Eignung hin gepruft werden. Anschliesend mussen diese Daten integriert und deren Heterogenitat auf verschiedenen Ebenen uberwunden werden.

Fusgangernavigationssysteme benutzen fur die Zielfuhrung idealerweise auffallige Objekte (Landmarken). Die Nachteile bisheriger Verfahren zur Gewinnung dieser Objekte konnen durch die Erfassungin alltaglichen Navigationssituationen mit Hilfe einer Smartphone-Applikation umgangen werden. Im Folgenden werden notige Vorarbeiten zur Umsetzung einer solchen Anwendung beschrieben: Zunachst mussen verschiedene Datenbestande auf ihre Eignung hin gepruft werden. Anschliesend mussen diese Daten integriert und deren Heterogenitat auf verschiedenen Ebenen uberwunden werden.

Fusgangernavigationssysteme benutzen fur die Zielfuhrung idealerweise auffallige Objekte (Landmarken). Die Nachteile bisheriger Verfahren zur Gewinnung dieser Objekte konnen durch die Erfassung in alltaglichen Navigationssituationen mithilfe einer Smartphone-Applikation umgangen werden. Im Folgenden werden notige Vorarbeiten zur Umsetzung einer solchen Anwendung beschrieben: Zunachst mussen verschiedene Datenbestande auf ihre Eignung hin gepruft werden. Anschliesend mussen diese Daten integriert und deren Heterogenitat auf verschiedenen Ebenen uberwunden werden.