Umeå University's logo

This paper describes the user evaluation of two strategies for guiding pedestrians. This work is based on a hands free, eyes free scenario explored by SpaceBook [1], using cheap Android phones. The first model described is a comparatively simple purely reactive model [2]. The second is using more involved techniques including prediction of future positions and scheduling [3]. The paper describes both models discussing identified problems with them. It then goes on to describe our evaluation procedure and the results of the evaluation.

We seek to test the hypothesis that text-to-speech(TTS) navigation systems can adequately guide pedestrians to unknown destinations in an unfamiliar city. Such systems bypass screenbased, multi-modal techniques and simply speak route following instructions incrementally into the pedestrian’s ear piece. Due to errors in GPS positioning, uncertainty of user heading, poor map quality and potential communication and processing latencies, this becomes a surprisingly challenging task. In our study, subjects are led on an unknown tour on the grounds of Ume˚a University. We evaluated both a human wizard controller as well as a simple decision-tree based controller and compared them to an ideal subject that knows the route. Results give support to our hypothesis that TTS-based navigation systems can adequately guide pedestrians. That said, our experiences point toward immediate and future improvements to make such systems more effective and agreeable. All the software and data behind this work will be open sourced to encourage confirmation, replication and, ultimately, improvement upon our results. This will soon be available for public download at http://janus-system.eu.

This paper presents an Android system to support eyes-free, hands-free navigation through a city. The system operates in two distinct modes: manual and automatic. In manual, a human operator sends text messages which are realized via TTS into the subject's earpiece. The operator sees the subject's GPS position on a map, hears the subject's speech, and sees a 1 fps movie taken from the subject's phone, worn as a necklace. In automatic mode, a programmed controller attempts to achieve the same guidance task as the human operator. We have fully built our manual system and have verified that it can be used to successfully guide pedestrians through a city. All activities are logged in the system into a single, large database state. We are building a series of automatic controllers which require us to confront a set of research challenges, some of which we briefly discuss in this paper. We plan to demonstrate our work live at NLDB.

Often we are interested to know what is 'near' and what is 'far' in spatial databases. For instance we would like a hotel 'near' the beach, but 'far' from the highway. It is not always obvious how to answer such nearness questions by reducing them to their crisp counterparts 'nearer' or 'nearest'. Thus we confront the vague and context-dependent relation of near (and far). Our approach follows a supervaluation tradition with a limited representation of context. The method is tractable, learnable and directly suitable for use in natural language interfaces to databases. The approach is based on logic programs supervaluated over a set of context-dependent threshold parameters. Given a set of rules with such unconstrained threshold parameters, a fixed parameter tractable algorithm finds a setting of parameters that are consistent with a training corpus of context-dependent descriptions of 'near' and 'far' in scenes. The results of this algorithm may then be compiled into view definitions which are accessed in real-time by natural language interfaces employing normal, non-exotic query answering mechanisms. 

This position paper makes a case for the need to predict pedestrian position and schedule communication acts in mobile navigation systems. In our work, carried out in the context of a voice guided city navigation system, we have found that improperly timed route instructions are a major cause of failure in guiding pedestrians in unknown environments. Furthermore, the need to communicate other information while guiding users on routes, as well as complications caused by network latencies, occurs often enough to require that we be able synchronize communication acts with user position as they follow a route. This has led us to focus our efforts on scheduling utterances to maximize route following success.

In this position paper we motivate this problem and present our initial approach and findings which should be of interest to others engaged in similar efforts in both the Geography and HCI communities.

We present an Android-based platform for incrementally presenting spoken route directions to guide pedestrians to destinations. Our approach makes heavy use of stored procedures and triggers in an underlying PostGIS spatial database. In fact most of the 'intelligence' of our prototype resides in database stored procedures and tables. As such it represents an example of a challenging real world case study for the use of persistent stored modules (PSM) in a complex mobility application. It also provides a platform to study performance tradeoffs for complex event processing over spatial data streams.

This paper demonstrates a way to build a natural language interface (NLI) over semantically rich data. Specifically we show this over the MusicBrainz domain, inspired by the second shared task of the QALD-1 workshop. Our approach uses the tool C-Phrase [4] to build an NLI over a set of views defined over the original MusicBrainz relational database. C-Phrase uses a limited variant of X-Bar theory [3] for syntax and tuple calculus for semantics. The C-Phrase authoring tool works over any domain and only the end configuration has to be redone for each new database covered – a task that does not require deep knowledge about linguistics and system internals. Working over the MusicBrainz domain was a challenge due to the size of the database – quite a lot of effort went into optimizing computation times and memory usage to manageable levels. This paper reports on this work and anticipates a live demonstration for querying by the public