Technical Program
The program is available as PDF file.
Invited Papers
Querying and Cleaning Uncertain Data
Dr. Reynold Cheng (University of Hong Kong)
Abstract The management of uncertainty in large databases has recently attracted tremendous research interest. Data uncertainty is inherent in many emerging and important applications, including location based services, wireless sensor networks, biometric and biological databases, and data stream applications. In these systems, it is important to manage data uncertainty carefully, in order to make correct decisions and provide high-quality services to users. To enable the development of these applications, uncertain database systems have been proposed. They consider data uncertainty as a "first-class citizen", and use generic data models to capture uncertainty, as well as provide query operators that return answers with statistical confidences. We summarize our work on uncertain databases in recent years. We explain how data uncertainty can be modeled, and present a classification of probabilistic queries (e.g., range query and nearest-neighbor query). We further study how probabilistic queries can be efficiently evaluated and indexed. We also highlight the issue of removing uncertainty under a stringent cleaning budget, with an attempt of generating high-quality probabilistic answers.
Dr. Reynold Cheng (University of Hong Kong)
Abstract The management of uncertainty in large databases has recently attracted tremendous research interest. Data uncertainty is inherent in many emerging and important applications, including location based services, wireless sensor networks, biometric and biological databases, and data stream applications. In these systems, it is important to manage data uncertainty carefully, in order to make correct decisions and provide high-quality services to users. To enable the development of these applications, uncertain database systems have been proposed. They consider data uncertainty as a "first-class citizen", and use generic data models to capture uncertainty, as well as provide query operators that return answers with statistical confidences. We summarize our work on uncertain databases in recent years. We explain how data uncertainty can be modeled, and present a classification of probabilistic queries (e.g., range query and nearest-neighbor query). We further study how probabilistic queries can be efficiently evaluated and indexed. We also highlight the issue of removing uncertainty under a stringent cleaning budget, with an attempt of generating high-quality probabilistic answers.
Context Quality and Privacy - Friends or Rivals?
Johann-Christoph Freytag (Humboldt-Universität zu Berlin)
Abstract As our world becomes more and more proliferated by sensors and mobile devices - often connected by wireless networks - there is the urging need to develop appropriate abstractions for application development and deployment. Those abstractions should shield applications from the physical properties of the devices thereby allowing applications to focus on information processing based on global conceptual views (of the world) in form of context models. For the benefit of the user many mobile devices and applications communicate with others to perform their tasks. Therefore, we see the need to give users some kind of control when and to which extend to allow applications to communicate with other (mobile) devices or applications. In particular, a user should be allowed to determine how much (s)he is willing to share personal (private)data with others when participating in such context aware infrastructures. That is, the user should have control over how much his/her personal data is accessed by or communicated to other systems if privacy is a concern to him/her. This paper will elaborate on the concern for privacy in location-aware systems by providing various examples that should highlight the complexity of such concerns. We show that privacy needs a well founded base for handling user requirements appropriately. Often, privacy is not a static property, but it is context sensitive thus increasing the overall complexity of managing privacy according to the user's expectations. Additionally, we argue that quality aspects in context model based systems should include and embed privacy protection and control mechanism as an integral part on all levels therefore increasing the usability of such systems from a user's point of view.
Johann-Christoph Freytag (Humboldt-Universität zu Berlin)
Abstract As our world becomes more and more proliferated by sensors and mobile devices - often connected by wireless networks - there is the urging need to develop appropriate abstractions for application development and deployment. Those abstractions should shield applications from the physical properties of the devices thereby allowing applications to focus on information processing based on global conceptual views (of the world) in form of context models. For the benefit of the user many mobile devices and applications communicate with others to perform their tasks. Therefore, we see the need to give users some kind of control when and to which extend to allow applications to communicate with other (mobile) devices or applications. In particular, a user should be allowed to determine how much (s)he is willing to share personal (private)data with others when participating in such context aware infrastructures. That is, the user should have control over how much his/her personal data is accessed by or communicated to other systems if privacy is a concern to him/her. This paper will elaborate on the concern for privacy in location-aware systems by providing various examples that should highlight the complexity of such concerns. We show that privacy needs a well founded base for handling user requirements appropriately. Often, privacy is not a static property, but it is context sensitive thus increasing the overall complexity of managing privacy according to the user's expectations. Additionally, we argue that quality aspects in context model based systems should include and embed privacy protection and control mechanism as an integral part on all levels therefore increasing the usability of such systems from a user's point of view.
Interacting with Context
Max Mühlhäuser (Technische Universität Darmstadt), Melanie Hartmann (Technische Universität Darmstadt)
Abstract Context is dodgy - just as the human computer user: hard to predict, erroneous, and probabilistic in nature. Linking the two together i.e. creating context-aware user interfaces (UIs) remains a great challenge in computer science since ubiquitous computing calls for lean, situated, and focused UIs that can be operated on the move or intertwined with primary tasks grabbing the user's attention. The paper reviews major categories of context that matter at the seam of humans and computers, emphasizing quality issues. Approaches to the marriage of context-awareness and user modeling are highlighted, including our own approach. Both sides of the coin are inspected: the improvement of UIs by means of quality attributed context information and, to a lesser extent, the challenge to convey context quality to the user as part of the interaction.
Max Mühlhäuser (Technische Universität Darmstadt), Melanie Hartmann (Technische Universität Darmstadt)
Abstract Context is dodgy - just as the human computer user: hard to predict, erroneous, and probabilistic in nature. Linking the two together i.e. creating context-aware user interfaces (UIs) remains a great challenge in computer science since ubiquitous computing calls for lean, situated, and focused UIs that can be operated on the move or intertwined with primary tasks grabbing the user's attention. The paper reviews major categories of context that matter at the seam of humans and computers, emphasizing quality issues. Approaches to the marriage of context-awareness and user modeling are highlighted, including our own approach. Both sides of the coin are inspected: the improvement of UIs by means of quality attributed context information and, to a lesser extent, the challenge to convey context quality to the user as part of the interaction.
Spatial Embedding and Spatial Context
Christopher Gold (University of Glamorgan)
Abstract A serious issue in urban 2D remote sensing is that even if you can identify linear features it is often difficult to combine these to form the object you want - the building. The classical example is of trees overhanging walls and roofs: it is often difficult to join the linear pieces together. For robot navigation, surface interpolation, GIS polygon "topology", etc., isolated 0D or 1D elements in 2D space are incomplete: they need to be fully embedded in 2D space in order to have a usable spatial context. We embed all our 0D and 1D entities in 2D space by means of the Voronoi diagram, giving a space-filling environment where spatial adjacency queries are straightforward. This has been an extremely difficult algorithmic problem. We show recent results. If we really want to move from exterior form to building functionality we must work with volumetric entities (rooms) embedded in 3D space. We thus need an adjacency model for 3D space, allowing queries concerning adjacency, access, etc. to be handled directly from the data structure, exactly as described for 2D space. We will show our recent results to handle this problem. We claim that an appropriate adjacency model greatly simplifies questions of spatial context of elements (such as walls) that may be extracted from raw data, allowing direct assembly of compound entities such as buildings. Relationships between compound objects provide solutions to building adjacency, robot navigation and related problems. If the spatial context can be stated clearly then other contextual issues may be greatly simplified.
Christopher Gold (University of Glamorgan)
Abstract A serious issue in urban 2D remote sensing is that even if you can identify linear features it is often difficult to combine these to form the object you want - the building. The classical example is of trees overhanging walls and roofs: it is often difficult to join the linear pieces together. For robot navigation, surface interpolation, GIS polygon "topology", etc., isolated 0D or 1D elements in 2D space are incomplete: they need to be fully embedded in 2D space in order to have a usable spatial context. We embed all our 0D and 1D entities in 2D space by means of the Voronoi diagram, giving a space-filling environment where spatial adjacency queries are straightforward. This has been an extremely difficult algorithmic problem. We show recent results. If we really want to move from exterior form to building functionality we must work with volumetric entities (rooms) embedded in 3D space. We thus need an adjacency model for 3D space, allowing queries concerning adjacency, access, etc. to be handled directly from the data structure, exactly as described for 2D space. We will show our recent results to handle this problem. We claim that an appropriate adjacency model greatly simplifies questions of spatial context of elements (such as walls) that may be extracted from raw data, allowing direct assembly of compound entities such as buildings. Relationships between compound objects provide solutions to building adjacency, robot navigation and related problems. If the spatial context can be stated clearly then other contextual issues may be greatly simplified.
The Quality of Geospatial Context
Michael Frank Goodchild (University of California, Santa Barbara)
Abstract The location of an event or feature on the Earth's surface can be used to discover information about the location's surroundings, and to gain insights into the conditions and processes that may affect or even cause the presence of the event or feature. Such reasoning lies at the heart of critical spatial thinking, and is increasingly implemented in tools such as geographic information systems and online Web mashups. But the quality of contextual information relies on accurate positions and descriptions. Over the past two decades substantial progress has been made on the theory and methods of geospatial uncertainty, but hard problems remain in several areas, including uncertainty visualization and propagation. Web 2.0 mechanisms are fostering the rapid growth of user-generated geospatial content, but raising issues of associated quality.
Michael Frank Goodchild (University of California, Santa Barbara)
Abstract The location of an event or feature on the Earth's surface can be used to discover information about the location's surroundings, and to gain insights into the conditions and processes that may affect or even cause the presence of the event or feature. Such reasoning lies at the heart of critical spatial thinking, and is increasingly implemented in tools such as geographic information systems and online Web mashups. But the quality of contextual information relies on accurate positions and descriptions. Over the past two decades substantial progress has been made on the theory and methods of geospatial uncertainty, but hard problems remain in several areas, including uncertainty visualization and propagation. Web 2.0 mechanisms are fostering the rapid growth of user-generated geospatial content, but raising issues of associated quality.
Research Papers
Establishing Similarity Across Multi-Granular Topological-Relation Ontologies
Matthew P. Dube (University of Maine), Max J. Egenhofer (University of Maine)
Abstract Within the Geospatial Semantic Web, selecting a different ontology for a spatial data set will enable that data's analysis in a different context. Analyses of multiple data sets, each based on a different ontology, require appropriate bridges across the ontologies. This paper focuses on establishing such a bridge across two topological-relation ontologies of different granularity - the standard eight detailed topological relations and five coarse topological relations. By mapping the conceptual neighborhood graphs onto a zonal representation, the different granularities are aligned spatially, yielding a reasoned approach to determining similarity values for the bridges across the two ontologies. A comparison with bridge lengths from an averaged model shows the better quality of zonal model.
Matthew P. Dube (University of Maine), Max J. Egenhofer (University of Maine)
Abstract Within the Geospatial Semantic Web, selecting a different ontology for a spatial data set will enable that data's analysis in a different context. Analyses of multiple data sets, each based on a different ontology, require appropriate bridges across the ontologies. This paper focuses on establishing such a bridge across two topological-relation ontologies of different granularity - the standard eight detailed topological relations and five coarse topological relations. By mapping the conceptual neighborhood graphs onto a zonal representation, the different granularities are aligned spatially, yielding a reasoned approach to determining similarity values for the bridges across the two ontologies. A comparison with bridge lengths from an averaged model shows the better quality of zonal model.
An Abstract Processing Model for the Quality of Context Data
Matthias Grossmann (University of Stuttgart), Nicola Hönle (University of Stuttgart), Carlos Lübbe (University of Stuttgart), Harald Weinschrott (University of Stuttgart)
Abstract Data quality can be relevant to many applications. Especially applications coping with sensor data cannot take a single sensor value for granted. Because of technical and physical restrictions each sensor reading is associated with an uncertainty. To improve quality, an application can combine data values from different sensors or, more generally, data providers. But as different data providers may have diverse opinions about a certain real world phenomenon, another issue arises: inconsistency. When handling data from different data providers, the application needs to consider their trustworthiness. This naturally introduces a third aspect of quality: trust. In this paper we propose a novel processing model integrating the three aspects of quality: uncertainty, inconsistency and trust.
Matthias Grossmann (University of Stuttgart), Nicola Hönle (University of Stuttgart), Carlos Lübbe (University of Stuttgart), Harald Weinschrott (University of Stuttgart)
Abstract Data quality can be relevant to many applications. Especially applications coping with sensor data cannot take a single sensor value for granted. Because of technical and physical restrictions each sensor reading is associated with an uncertainty. To improve quality, an application can combine data values from different sensors or, more generally, data providers. But as different data providers may have diverse opinions about a certain real world phenomenon, another issue arises: inconsistency. When handling data from different data providers, the application needs to consider their trustworthiness. This naturally introduces a third aspect of quality: trust. In this paper we propose a novel processing model integrating the three aspects of quality: uncertainty, inconsistency and trust.
A Probabilistic Filter Protocol for Continuous Queries
Jinchuan Chen (Hong Kong University), Reynold Cheng (Hong Kong University), Yinuo Zhang (Hong Kong University), Jian Jin (Hong Kong University)
Abstract Pervasive applications, such as location-based services and natural habitat monitoring, have attracted plenty of research interest. These applications make use of a large number of remote positioning devices like Global Positioning System (GPS) for collecting users' physical locations. Generally, these devices have battery power limitation. They also cannot report very accurate position values. In this paper, we consider the evaluation of a long-standing (or continuous) query over inaccurate location data collected from positioning devices. Our goal is to develop an energy-efficient protocol, which provides some degree of confidence on the query answers evaluated on imperfect data. In particular, we propose the probabilistic filter, which governs GPS devices to decide upon whether location values collected should be reported to the server. We further discuss how these filters can be developed. This scheme reduces the cost of transmitting location updates, and hence the energy spent by the GPS devices. It also allows some portion of query processing to be deployed to the devices, thereby alleviating the processing burden of the server.
Jinchuan Chen (Hong Kong University), Reynold Cheng (Hong Kong University), Yinuo Zhang (Hong Kong University), Jian Jin (Hong Kong University)
Abstract Pervasive applications, such as location-based services and natural habitat monitoring, have attracted plenty of research interest. These applications make use of a large number of remote positioning devices like Global Positioning System (GPS) for collecting users' physical locations. Generally, these devices have battery power limitation. They also cannot report very accurate position values. In this paper, we consider the evaluation of a long-standing (or continuous) query over inaccurate location data collected from positioning devices. Our goal is to develop an energy-efficient protocol, which provides some degree of confidence on the query answers evaluated on imperfect data. In particular, we propose the probabilistic filter, which governs GPS devices to decide upon whether location values collected should be reported to the server. We further discuss how these filters can be developed. This scheme reduces the cost of transmitting location updates, and hence the energy spent by the GPS devices. It also allows some portion of query processing to be deployed to the devices, thereby alleviating the processing burden of the server.
Presentation and Evaluation of Inconsistencies in Multiply Represented 3D Building Models
Michael Peter (University of Stuttgart)
Abstract Open architectures demand for a federation of data from different context providers, which nearly always will be inconsistent to a certain degree. We present an approach for the evaluation and presentation of inconsistencies in multiply represented 3D building models and provide means for the minimization of ground plan inconsistencies. The presented approaches are tested using differently detailed models from various sources.
Michael Peter (University of Stuttgart)
Abstract Open architectures demand for a federation of data from different context providers, which nearly always will be inconsistent to a certain degree. We present an approach for the evaluation and presentation of inconsistencies in multiply represented 3D building models and provide means for the minimization of ground plan inconsistencies. The presented approaches are tested using differently detailed models from various sources.
Quality dependent reconstruction of building façades
Susanne Becker (University of Stuttgart), Norbert Haala (University of Stuttgart)
Abstract The paper describes an approach for the quality dependent reconstruction of building façades using 3D point clouds from mobile terrestrial laser scanning. Due to different look angles, such measurements frequently suffer from different point densities at the respective building façades. In order to support the interpretation at areas, where no or only limited LiDAR measurements are available, a quality dependent processing is implemented. First, façades are reconstructed at areas of sufficient LiDAR point availability. Based on this reconstruction, rules are derived automatically, which together with the respective façade elements constitute a so-called façade grammar. It holds all the information which is necessary to reconstruct façades in the style of the given building. In our quality dependent approach, this grammar is used as knowledge base for the verification of a façade model reconstructed at areas of limited sensor data quality. Additionally, it is applied for the generation of synthetic façades where no LiDAR measurement is available.
Susanne Becker (University of Stuttgart), Norbert Haala (University of Stuttgart)
Abstract The paper describes an approach for the quality dependent reconstruction of building façades using 3D point clouds from mobile terrestrial laser scanning. Due to different look angles, such measurements frequently suffer from different point densities at the respective building façades. In order to support the interpretation at areas, where no or only limited LiDAR measurements are available, a quality dependent processing is implemented. First, façades are reconstructed at areas of sufficient LiDAR point availability. Based on this reconstruction, rules are derived automatically, which together with the respective façade elements constitute a so-called façade grammar. It holds all the information which is necessary to reconstruct façades in the style of the given building. In our quality dependent approach, this grammar is used as knowledge base for the verification of a façade model reconstructed at areas of limited sensor data quality. Additionally, it is applied for the generation of synthetic façades where no LiDAR measurement is available.
UDS: Sustaining Quality of Context using Uninterruptible Data Supply System
Naoya Namatame (Keio University), Jin Nakazawa (Keio University), Kazunori Takashio (Keio University), Hideyuki Tokuda (Keio University)
Abstract Context mining algorithms from sensor data have been researched and successful results have been shown. However, since these existing works are focused on improving the accuracy of context mining, they are established on the assumption that they can acquire a complete set of necessary data. Therefore, the context mining algorithms do not work sufficiently since the data drops easily in the reality. In this paper, to cope with this problem, we propose a middleware named UDS (Uninterruptible Data Supply System). The system compensates the missing data, creates virtually complete dataset and provides upper layer applications. Applications operating over UDS can work sufficiently with some data actually missing. We have defined two types of characteristic data deficit patterns and created a robust model for both patterns utilizing Bayesian Network. In the evaluation, we show UDS can sustain the quality of context over 80% with 40% data missing.
Naoya Namatame (Keio University), Jin Nakazawa (Keio University), Kazunori Takashio (Keio University), Hideyuki Tokuda (Keio University)
Abstract Context mining algorithms from sensor data have been researched and successful results have been shown. However, since these existing works are focused on improving the accuracy of context mining, they are established on the assumption that they can acquire a complete set of necessary data. Therefore, the context mining algorithms do not work sufficiently since the data drops easily in the reality. In this paper, to cope with this problem, we propose a middleware named UDS (Uninterruptible Data Supply System). The system compensates the missing data, creates virtually complete dataset and provides upper layer applications. Applications operating over UDS can work sufficiently with some data actually missing. We have defined two types of characteristic data deficit patterns and created a robust model for both patterns utilizing Bayesian Network. In the evaluation, we show UDS can sustain the quality of context over 80% with 40% data missing.
On a Generic Uncertainty Model for Position Information
Ralph Lange (University of Stuttgart), Harald Weinschrott (University of Stuttgart) Lars Geiger (University of Stuttgart), Andre Blessing (University of Stuttgart), Frank Dürr (University of Stuttgart), Kurt Rothermel (University of Stuttgart), Hinrich Schütze (University of Stuttgart)
Abstract Position information of moving as well as stationary objects is generally subject to uncertainties due to inherent measuring errors of positioning technologies, explicit tolerances of position update protocols, and approximations by interpolation algorithms. There exist a variety of approaches for specifying these uncertainties by mathematical uncertainty models such as tolerance regions or the Dilution of Precision (DOP) values of GPS. In this paper we propose a principled generic uncertainty model that integrates the different approaches and derive a comprehensive query interface for processing spatial queries on uncertain position information of different sources based on this model. Finally, we show how to implement our approach with prevalent existing uncertainty models.
Ralph Lange (University of Stuttgart), Harald Weinschrott (University of Stuttgart) Lars Geiger (University of Stuttgart), Andre Blessing (University of Stuttgart), Frank Dürr (University of Stuttgart), Kurt Rothermel (University of Stuttgart), Hinrich Schütze (University of Stuttgart)
Abstract Position information of moving as well as stationary objects is generally subject to uncertainties due to inherent measuring errors of positioning technologies, explicit tolerances of position update protocols, and approximations by interpolation algorithms. There exist a variety of approaches for specifying these uncertainties by mathematical uncertainty models such as tolerance regions or the Dilution of Precision (DOP) values of GPS. In this paper we propose a principled generic uncertainty model that integrates the different approaches and derive a comprehensive query interface for processing spatial queries on uncertain position information of different sources based on this model. Finally, we show how to implement our approach with prevalent existing uncertainty models.
A Context Quality Model to Support Transparent Reasoning with Uncertain Context
Susan McKeever (University College Dublin), Juan Ye (University College Dublin), Lorcan Coyle (University College Dublin), Simon Dobson (University College Dublin)
Abstract Much research on context quality in context-aware systems divides into two strands: (1) the qualitative identification of quality measures and (2) the use of uncertain reasoning techniques. In this paper, we combine these two strands, exploring the problem of how to identify and propagate quality through the different context layers in order to support the context reasoning process. We present a generalised, structured context quality model that supports aggregation of quality from sensor up to situation level. Our model supports reasoning processes that explicitly aggregate context quality, by enabling the identification and quantification of appropriate quality parameters. We demonstrate the efficacy of our model using an experimental sensor data set, gaining a significant improvement in situation recognition for our voting based reasoning algorithm.
Susan McKeever (University College Dublin), Juan Ye (University College Dublin), Lorcan Coyle (University College Dublin), Simon Dobson (University College Dublin)
Abstract Much research on context quality in context-aware systems divides into two strands: (1) the qualitative identification of quality measures and (2) the use of uncertain reasoning techniques. In this paper, we combine these two strands, exploring the problem of how to identify and propagate quality through the different context layers in order to support the context reasoning process. We present a generalised, structured context quality model that supports aggregation of quality from sensor up to situation level. Our model supports reasoning processes that explicitly aggregate context quality, by enabling the identification and quantification of appropriate quality parameters. We demonstrate the efficacy of our model using an experimental sensor data set, gaining a significant improvement in situation recognition for our voting based reasoning algorithm.
Using Quality of Context to Resolve Conflicts in Context-Aware Systems
Atif Manzoor (Vienna University of Technology)
Hong-Linh Truong (Vienna University of Technology), Schahram Dustdar (Vienna University of Technology)
Abstract Context-aware systems in mobile and pervasive environments face many conflicting situations while collecting sensor data, processing sensor data to extract consistent and coherent high level context information, and disseminating that context information to assist in making decisions to adapt to the continuously evolving situations without diverting human attention. These conflicting situations pose stern challenges to the design and development of context-aware systems by making it extremely complicated and error-prone. Quality of Context parameters can be used to cope with these challenges. In this paper, we discuss the conflicting situations that a context-aware system can face at different layers of its conceptual design and present the conflict resolving policies that are defined on the basis of the Quality of Context parameters. We also illustrate how these policies can be used in different conflicting situations to improve the performance and effectiveness of context-aware systems.
Atif Manzoor (Vienna University of Technology)
Hong-Linh Truong (Vienna University of Technology), Schahram Dustdar (Vienna University of Technology)
Abstract Context-aware systems in mobile and pervasive environments face many conflicting situations while collecting sensor data, processing sensor data to extract consistent and coherent high level context information, and disseminating that context information to assist in making decisions to adapt to the continuously evolving situations without diverting human attention. These conflicting situations pose stern challenges to the design and development of context-aware systems by making it extremely complicated and error-prone. Quality of Context parameters can be used to cope with these challenges. In this paper, we discuss the conflicting situations that a context-aware system can face at different layers of its conceptual design and present the conflict resolving policies that are defined on the basis of the Quality of Context parameters. We also illustrate how these policies can be used in different conflicting situations to improve the performance and effectiveness of context-aware systems.
A Framework for Quality of Context Management
Zied Abid (Institut Télécom), Sophie Chabridon (Institut Télécom), Denis Conan (Institut Télécom)
Abstract Context-aware computing has to deal with a huge amount of context data. Taking into account the quality of these data becomes a corner stone of an efficient context management solution. Information on the quality of context helps taking appropriate decisions and allows to identify uncertain context information saving processing time for deriving a pertinent description of the observed phenomenon. This paper presents a work in progress for integrating Quality of Context in COSMOS (COntext entitieS coMpositiOn and Sharing), a component-based framework for managing context data in ubiquitous environments, and illustrates it throughout the example of the composition of context information to implement a network connectivity vs energy adaptation situation.
Zied Abid (Institut Télécom), Sophie Chabridon (Institut Télécom), Denis Conan (Institut Télécom)
Abstract Context-aware computing has to deal with a huge amount of context data. Taking into account the quality of these data becomes a corner stone of an efficient context management solution. Information on the quality of context helps taking appropriate decisions and allows to identify uncertain context information saving processing time for deriving a pertinent description of the observed phenomenon. This paper presents a work in progress for integrating Quality of Context in COSMOS (COntext entitieS coMpositiOn and Sharing), a component-based framework for managing context data in ubiquitous environments, and illustrates it throughout the example of the composition of context information to implement a network connectivity vs energy adaptation situation.
Bringing Quality of Context into Wearable Human Activity Recognition Systems
Claudia Villalonga (ETH Zurich), Daniel Roggen (ETH Zurich), Clemens Lombriser (ETH Zurich), Piero Zappi (ETH Zurich), Gerhard Tröster (ETH Zurich)
Abstract Quality of Context (QoC) in context-aware computing improves reasoning and decision making. Activity recognition in wearable computing enables context-aware assistance. Wearable systems must include QoC to participate in context processing frameworks common in large ambient intelligence environments. However, QoC is not specifically defined in that domain. QoC models allowing activity recognition system reconfiguration to achieve a desired context quality are also missing. Here we identify the recognized dimensions of QoC and the performance metrics in activity recognition systems. We discuss how the latter maps on the former and provide guidelines to include QoC in activity recognition systems. On the basis of gesture recognition in a car manufacturing case study, we illustrate the signification of QoC and we present modeling abstractions to reconfigure an activity recognition system to achieve a desired QoC.
Claudia Villalonga (ETH Zurich), Daniel Roggen (ETH Zurich), Clemens Lombriser (ETH Zurich), Piero Zappi (ETH Zurich), Gerhard Tröster (ETH Zurich)
Abstract Quality of Context (QoC) in context-aware computing improves reasoning and decision making. Activity recognition in wearable computing enables context-aware assistance. Wearable systems must include QoC to participate in context processing frameworks common in large ambient intelligence environments. However, QoC is not specifically defined in that domain. QoC models allowing activity recognition system reconfiguration to achieve a desired context quality are also missing. Here we identify the recognized dimensions of QoC and the performance metrics in activity recognition systems. We discuss how the latter maps on the former and provide guidelines to include QoC in activity recognition systems. On the basis of gesture recognition in a car manufacturing case study, we illustrate the signification of QoC and we present modeling abstractions to reconfigure an activity recognition system to achieve a desired QoC.