Papers by Kai-florian Richter
We present a dialogue system that automatically generates indoor route instructions in German whe... more We present a dialogue system that automatically generates indoor route instructions in German when asked about locations, using text-based natural language input and output. The challenging task in this system is to provide the user with a compact set of accurate and comprehensible instructions. We describe our approach based on high-level instructions. The system is described with four main modules: natural language understanding, dialogue management, route instruction generation and natural language generation. We report an evaluation with users unfamiliar with the system -using the PARADISE evaluation frameworkin a real environment and naturalistic setting. We present results with high user satisfaction, and discuss future directions for enhancing this kind of system with more sophisticated and intuitive interaction.
People communicate about locations using place descriptions. Despite the growth of mobile locatio... more People communicate about locations using place descriptions. Despite the growth of mobile location- and context-aware applications, the automatic interpretation of place descriptions remains a challenge. Currently no software tools exist that are capable of understanding complex verbal spatial language. This paper explores a corpus of place descriptions collected through crowdsourcing mechanisms within a mobile game. It introduces a general classification scheme to annotate place descriptions according to different characteristic parameters and uses this scheme to demonstrate the existence of certain clusters of prevalent types of place descriptions in human communication. Research outcomes contribute to the common understanding of the way people refer to places, which is essential to support the development of intelligent tools and location based technologies.
Lecture Notes in Computer Science, 2000
Shape simplification in map-like representations is used for two reasons: either to abstract from... more Shape simplification in map-like representations is used for two reasons: either to abstract from irrelevant detail to reduce a map user's cognitive load, or to simplify information when a map of a smaller scale is derived from a detailed reference map. We present a method for abstracting simplified cartographic representations from more accurate spatial data. First, the employed method of discrete curve evolution developed for simplifying perceptual shape characteristics is explained. Specific problems of applying the method to cartographic data are elaborated. An algorithm is presented, which on the one hand simplifies spatial data up to a degree of abstraction intended by the user; and which on the other hand does not violate local spatial ordering between (elements of) cartographic entities, since local arrangement of entities is assumed to be an important spatial knowledge characteristic. The operation of the implemented method is demonstrated using two different examples of cartographic data.
We present an approach to mobile robot guidance. The proposed architecture grants an abstract int... more We present an approach to mobile robot guidance. The proposed architecture grants an abstract interface to robot navigation allowing to bridge from perception to high level control. The approach is based upon a comprehensive map representing metric, configurational, and topological knowledge. Robot instruction and localization is dealt with by communicating and reasoning about cyclic ordering information of shape-features.
Lecture Notes in Computer Science, 2014
Humans engage in wayfinding many times a day. We try to find our way in urban environments when w... more Humans engage in wayfinding many times a day. We try to find our way in urban environments when walking towards our work places or when visiting a city as tourists. In order to reach the targeted destination, we have to make a series of wayfinding decisions of varying complexity. Previous research has focused on classifying the complexity of these wayfinding decisions, primarily looking at the complexity of the decision point itself (e.g., the number of possible routes or branches). In this paper, we proceed one step further by incorporating the user, instructions, and environmental factors into a model that assesses the complexity of a wayfinding decision. We constructed and evaluated three models using data collected from an outdoor wayfinding study. Our results suggest that additional factors approximate the complexity of a wayfinding decision better than the simple model using only the number of branches as a criterion.
Lecture Notes in Computer Science, 2005
Wayfinding, i.e. getting from some origin to a destination, is one of the prime everyday problems... more Wayfinding, i.e. getting from some origin to a destination, is one of the prime everyday problems humans encounter. It has received a lot of attention in research and many (commercial) systems propose assistance in this task. We present an approach to route directions based on the idea to adapt route directions to route and environment's characteristics. The lack of such an adaptation is a major drawback of existing systems. Our approach is based on an information-and representation-theoretic analysis of routes and takes into account findings of behavioral research. The resulting systematics is the framework for the optimization process. We discuss the consequences of using an optimization process for generating route directions and outline its algorithmic realization.
Lecture Notes in Computer Science, 2007
Cognitive agents use different strategies to identify relevant spatial information in communicati... more Cognitive agents use different strategies to identify relevant spatial information in communication. The chosen strategy depends on the agents' conceptualization of the spatial situation at hand. This situation is determined by structural and functional aspects that are induced by the environment and the actions performed or intended therein. In this paper, we are interested in conceptualizations in the context of route directions. We focus on the meaning of prepositions used to characterize movements (actions) in spatially constrained systems such as street networks. We report on different strategies employed by people to disambiguate turning actions at intersections and demonstrate how these can be reflected in automatically generated route directions, again concentrating on the assignment of prepositions for anchoring movement. Including methods that focus on the most successful strategies people use in computational systems is a prerequisite for route directions that respect for human conceptualizations of spatial situations and that become, thus, cognitively ergonomic route directions.
RoadPlex: A Mobile VGI Game to Collect and Validate Data for POIs
ISPRS Annals of Photogrammetry, Remote Sensing and Spatial Information Sciences, 2014
ABSTRACT Because of increasing popularity of smart phones equipped with GPS sensors, more volunte... more ABSTRACT Because of increasing popularity of smart phones equipped with GPS sensors, more volunteers are expected to join VGI (Volunteered Geographic Information) activities and therefore more positional data will be collected in shorter time. Current statistics from open databases such as OpenStreetMap reveal that although there has been an exponential growth in the number of contributed POIs (Points of Interest), the lack of detailed attribute information is immediately visible. The process of adding attribute information to VGI databases is usually considered as a boring task and it is believed that contributors do not experience a similar level of satisfaction when they add such detailed information compared to tasks like adding new roads or copying building boundaries from satellite imageries. In other crowdsourcing projects, different approaches are taken for engaging contributors in problem solving by ‘hiding’ the tasks behind a game. In the literature, this concept is known as “gamification” or “games with purpose” which encapsulate the idea of entertaining contributors while they are completing a particular defined task. The same concept is used to design a mobile application called “RoadPlex” which aims to collect attribute information for POIs. The increased number of contributions in the past few months confirms that the design characteristics and the methodology of the game are appealing to players. Such growth enables us to evaluate the quality of the generated data through mining the database of answered questions. This paper discusses some contribution results and emphasises the importance of using gamification concepts in the domain of VGI.
Supporting user-specific views via multidimensional trees
CHI '01 extended abstracts on Human factors in computing systems - CHI '01, 2001
Abstract Large sets of information need to be structured in order to be usable. Users often build... more Abstract Large sets of information need to be structured in order to be usable. Users often build hierarchical representations of information that are both user-and task-specific. Traditional structuring techniques used in information systems often fail to support these hierarchies. We try to support the use of individual concepts of the information space with a general structure. We also try to minimize the number of forced decisions in the user's decision tree.
Spatial Cognition, 2010
We present a learning approach for efficiently inducing adaptive behaviour of route instructions.... more We present a learning approach for efficiently inducing adaptive behaviour of route instructions. For such a purpose we propose a two-stage approach to learn a hierarchy of wayfinding strategies using hierarchical reinforcement learning. Whilst the first stage learns low-level behaviour, the second stage focuses on learning high-level behaviour. In our proposed approach, only the latter is to be applied at
Over the last ten years, various standards for the representation of generated wayfinding instruc... more Over the last ten years, various standards for the representation of generated wayfinding instructions have been proposed in the Geographic Information Science (GIS) literature . The more advanced of these standards specify wayfinding information types which may be rendered to users at various levels of granularity, and in multiple modalities. While necessary and sufficient for the present generation of GI services, the integration of wayfinding strategies with more general cognitive and mobility assistance systems will require that such descriptive standards be anchored within, or at least related to, other elements of a cognitive assistance framework. To progress us towards this goal, in this talk we analyse the ontological foundations of multimodal wayfinding description standards and relate these to the state of the art in both linguistic and conceptual ontologies. Linguistic ontologies provide an interface with rich natural language processing resources, while conceptual ontologies provide a means by which the wayfinding structures can be directly related to other aspects of spatial modelling -including the spatial structures against which route instructions are generated. We report on our progress in this area, and on the relationship between wayfinding description standards and the needs of the route interpretation community. Our analysis focuses in particular on the notion of action, place, and direction as used within wayfinding description standards. The analysis presented is part of ongoing work into the integration of the state of the art in intelligent wayfinding description methods with ontologically well motivated cognitive dialogue systems.
THE COGNITIVE APPROACH TO MODELING ENVIRONMENTS
... 37 SESSION 3: REPRESENTATION OF ENVIRONMENTS Cognitive Structure, Urban SymbolicOrder and Lan... more ... 37 SESSION 3: REPRESENTATION OF ENVIRONMENTS Cognitive Structure, Urban SymbolicOrder and Landmark Detection Ana Paula Neto de Faria & Romulo Krafta..... 41 On Modeling ...
Technologies for capturing large amounts of real-time and high-detail data about the environment ... more Technologies for capturing large amounts of real-time and high-detail data about the environment have advanced rapidly; our ability to use this data for understanding the monitored settings for decision-making has not. Visual analytics, creating suitable tools and interfaces that combine computational powers with the human's capabilities for visual sense making, is a promising approach. Geosensor networks monitor a range of different complex environmental settings, collecting heterogeneous data at different spatial and temporal scales. Similarly domain experts with specific preferences and requirements use the collected data. Additionally, long-term monitoring networks may aim to increase sensor node longevity by minimizing storage and communication load. Based on these aspects, four key challenges for the extraction of knowledge about environmental objects and events from geosensor data are identified: dynamics and uncertainty of the continuous stream of recorded data; different scales in data collection but also data analysis at a range of aggregation levels; decentralized data processing and storage; and evaluation of the effectiveness, efficiency and completeness of implemented decentralized visual analytics approaches.
The impact of classification approaches on the detection of hierarchies in place descriptions
ABSTRACT The chapter investigates the identification of hierarchical structures in place descript... more ABSTRACT The chapter investigates the identification of hierarchical structures in place descriptions. Different approaches to classify spatial granularity will be compared and applied to a corpus of human place descriptions. Results show how hierarchical structures as well as deviations depend on the respective classifications. They further indicate certain difficulties in developing a suitable classification of spatial references. Findings contribute to the understanding of human spatial language, and thus the development of flexible mechanisms for their interpretation and integration in location-based systems.
Place descriptions are a predominant means of human spatial communication. Their automated interp... more Place descriptions are a predominant means of human spatial communication. Their automated interpretation still poses a challenge for geospatial services. This paper explores one issue of this interpretation process: determining the level of granularity to which a localization of a described place is possible. Knowing this finest possible level of granularity supports resolving place descriptions, for example, in geographic information retrieval. In particular, the focus is on integrating spatial relations into this process. To this end, a mechanistic procedure for determining the level of granularity is proposed and applied to a place descriptions corpus. Feasibility of the procedure is evaluated in a comparison of place descriptions with people's self-reported position on a map. Findings show that the procedure delivers generally good results in agreement with the corresponding map locations.
Decentralized evacuation management
ABSTRACT Evacuation is an urgent measure of disaster response. It requires route planning by many... more ABSTRACT Evacuation is an urgent measure of disaster response. It requires route planning by many individual agents under circumstances that include the agents’ limited knowledge of their environment, an unknown impact of the disaster on the environment, and potentially destroyed, blocked, congested or lacking communication infrastructure. These circumstances impede any straightforward approach to evacuation planning. Neither are individuals able to determine an optimal route, nor is a centralized service able to determine or communicate routes to individuals. This paper suggests and investigates a novel paradigm for evacuation management: decentralized planning based on sharing local knowledge in a peer-to-peer manner. The paradigm is independent from external communication infrastructure, adapts to dynamic disasters, and turns out to be in many scenarios as successful as centralized management. Results have implications for disaster management practice, and cooperative intelligent transport in general.
Locating place names from place descriptions
ABSTRACT In this paper, we review the current literature on geographic information retrieval base... more ABSTRACT In this paper, we review the current literature on geographic information retrieval based on place names. We focus on the positional uncertainties and the extent of vagueness frequently associated with place names in linguistic place descriptions and on the differences between common users’ perception and the way the geographic information services interpret place names. We argue that, despite some notable efforts from the scientific community, geographic information services still cannot unambiguously recognize and sufficiently perform spatial reasoning with place names used in linguistic expressions. We focus on three interrelated research areas: 1 the use of place names in gazetteers, 2 the use of formal models to reason with spatial relations and with the spatial extent of place names in linguistic place descriptions, and 3 Web-harvesting and crowd-sourcing techniques for identifying place names and their spatial extension from public and volunteer sources, such as social networks and photo-sharing sites. We identify some opportunities for synthesizing existing approaches that would expedite the process of intelligent communication about place names between services and users. We discuss the shortcomings of the current state of affairs in locating place names from place descriptions and identify new areas of importance for future research.
Spatial interpretations of preposition "at
Proceedings of the 1st ACM SIGSPATIAL International Workshop on Crowdsourced and Volunteered Geographic Information - GEOCROWD '12, 2012
ABSTRACT The current keyword- and substring matches-based retrieval methods most search engines r... more ABSTRACT The current keyword- and substring matches-based retrieval methods most search engines rely on to answer spatial queries ignore the more specific interpretations of spatial relations. Moreover, the use of the general preposition "at" in natural language queries results in underspecified locations. This paper examines the use of "at" in a set of crowdsourced place descriptions and develops a methodology for interpreting "at" as one of the more location-specific, closely related spatial prepositions "in", "on", or "by". The application of suggested schemas in the paper enables the interpretation of "at" according to the granularity level and classification type of the spatial feature it refers to. The crowdsourced results show that most people use "at" to locate themselves either in relation to buildings, or to (naturally or artificially) bounded outdoor areas (at street level). When used with building level reference features, "at" is more likely to be interpreted as "in" the feature (mostly indoors) and less so "by" it. For reference features at street level, "at" is more likely to be interpreted as "in" the reference feature's region (within a bounded outdoor area) and less so "on" it (e.g., a water surface). The results indicate it is possible to use the proposed methodology for enabling search engines to better rank the results returned to natural language spatial queries by appropriately interpreting "at". The paper's research outcomes are an example of the use of crowdsourced information for improving the interaction between users and services.
2008 12th International Conference on Computer Supported Cooperative Work in Design, 2008
Design is a complex process often described as a collaboration between a designer and herself. In... more Design is a complex process often described as a collaboration between a designer and herself. In architectural design, a designer must consider the selection of a subset of actions from a set of possible actions while accounting for consequences on the overall design. Computer-based design tools enable humans to operate more efficiently in this process. We present CoSyCAD, a program that can be used to assist architects in the layout of a floor plan and to simultaneously analyze the cognitive complexity of routes through an indoor environment, thereby enabling direct feedback on a layout's usability. We provide a scenario that utilizes the program.
Spatial Cognition and Computation, 2010
In unfamiliar environments, people need assistance to find their way. One predominant form of suc... more In unfamiliar environments, people need assistance to find their way. One predominant form of such assistance is maps. In constructing these maps, there is a conflict between concentrating on the essential information for wayfinding, namely the route, and providing overview information of the environment. The former eases information extraction by reducing visual clutter, the latter allows for reorientation in the
Uploads
Papers by Kai-florian Richter