Catálogo de publicaciones - libros

The knowledge of the position of a user is valuable for a broad range of applications in the field of pervasive computing. Different techniques have been developed to cope with the problem of uncertainty, noisy sensors, and sensor fusion.

In this paper we present a method, which is efficient in time- and space-complexity, and that provides a high scalability for in- and outdoor-positioning. The so-called geo referenced dynamic Bayesian networks enable the calculation of a user’s position on his own small hand-held device (e.g., Pocket PC) without a connection to an external server. Thus, privacy issues are considered and completely in the hand of the user.

Research in advanced context-aware systems has clearly shown a need to capture the inherent uncertainty in the physical world, especially in human behavior. Modelling approaches that employ the concept of probability, especially in combination with Bayesian methods, are promising candidates to solve the pending problems. This paper analyzes the requirements for such models in order to enable user-friendly, adaptive and especially scalable operation of context-aware systems. It is conjectured that a successful system may not only use Bayesian techniques to infer probabilities from known probability tables but learn, i.e. estimate the probabilities in these tables by observing user behavior.

In this paper we present a context-aware collaborative filtering system that predicts a user’s preference in different context situations based on past experiences. We extend collaborative filtering techniques so that what other like-minded users have done in similar context can be used to predict a user’s preference towards an activity in the current context. Such a system can help predict the user’s behavior in different situations without the user actively defining it. For example, it could recommend activities customized for Bob for the given weather, location, and traveling companion(s), based on what other people like Bob have done in similar context.

In this paper, we introduce a compact system for fusing location data with data from simple, low-cost, non-location sensors to infer a user’s place and situational context. Specifically, the system senses location with a GSM cell phone and a WiFi-enabled mobile device (each running Place Lab), and collects additional sensor data using a 2” x 1” sensor board that contains a set of common sensors (e.g. accelerometers, barometric pressure sensors) and is attached to the mobile device. Our chief contribution is a multi-sensor system design that provides indoor-outdoor location information, and which models the capabilities and form factor of future cell phones. With two basic examples, we demonstrate that even using fairly primitive sensor processing and fusion algorithms we can leverage the synergy between our location and non-location sensors to unlock new possibilities for mobile context inference. We conclude by discussing directions for future work.

The paper describes a method that allows us to derive the location of an acceleration sensor placed on the user’s body solely based on the sensor’s signal. The approach described here constitutes a first step in our work towards the use of sensors integrated in standard appliances and accessories carried by the user for complex context recognition. It is also motivated by the fact that device location is an important context (e.g. glasses being worn vs. glasses in a jacket pocket). Our method uses a (sensor) location and orientation invariant algorithm to identify time periods where the user is walking and then leverages the specific characteristics of walking motion to determine the location of the body-worn sensor.

In the paper we outline the relevance of sensor location recognition for appliance based context awareness and then describe the details of the method. Finally, we present the results of an experimental study with six subjects and 90 walking sections spread over several hours indicating that reliable recognition is feasible. The results are in the low nineties for frame by frame recognition and reach 100% for the more relevant event based case.

Context information is used by pervasive networking and context-aware programs to adapt intelligently to different environments and user tasks. As the context information is potentially sensitive, it is often necessary to provide privacy protection mechanisms for users. These mechanisms are intended to prevent breaches of user privacy through unauthorised context disclosure. To be effective, such mechanisms should not only support user specified context disclosure rules, but also the disclosure of context at different granularities. In this paper we describe a new obfuscation mechanism that can adjust the granularity of different types of context information to meet disclosure requirements stated by the owner of the context information. These requirements are specified using a preference model we developed previously and have since extended to provide granularity control. The obfuscation process is supported by our novel use of ontological descriptions that capture the granularity relationship between instances of an object type.

Anonymity and location privacy in mobile and pervasive environments has been receiving increasing attention during the last few years, and several mechanisms and architectures have been proposed to prevent “big brother” phenomena. In this paper we present a novel architecture to shield the location of a mobile user and preserve the anonymity on the service delivery. This architecture, called “Share the Secret – STS”, uses totally un-trusted entities to distribute portions of anonymous location information, and authorizes other entities to combine these portions and derive the location of a user. STS simply divides the secret, and as a lightweight scheme it can be applied to network of nodes illustrating low processing and computational power, such as nodes of an ad-hoc network, smart gadgets and sensors.

In this paper we present an approach for exploiting knowledge about features in the real world in order to compute visibility of buildings. This is performed with the awareness of the inconsistencies and lack of accuracy in both mapping technology and GPS positioning in urban spaces. Electronic tourist guide systems typically recommend locations and sometimes provide navigation information. We have augmented this system to exploit visibility knowledge about neighbouring physical features.

We describe our approach of introducing context-awareness into everyday applications to make them more easy-to-use. The approach aims in shortening both the learning curve when introducing new technology to end-users and prototype development time, as well as results in more reliable prototypes. Moreover, we expect that the approach yields better quality user test results. To demonstrate the approach, we have employed context-based availability inference to automatically update the availability of IBM Lotus Sametime Everyplace users. This is likely to result in more reliable availability information and to make the application easier to use. Context inference is done using information from Lotus Notes Calendar and WLAN positioning technology.

In wireless networks the quality of service (QoS) delivered to an end user will be a complex function of location and time. “QoS Seeker” is a new system which informs a user what location in the wireless network he should move to in order deliver the required QoS for his real-time applications. At the heart of QoS Seeker is a construct known as a “QoS Map” – which is the value of a specific QoS metric as a function of space.If any significant temporal trends are present in the wireless network, then the current QoS Map will be statistically different from a future QoS Map. In this report we investigate the use of adaptive linear filters as a means to predict future QoS Maps from historical QoS Maps. By using the received signal strength (RSS) as the QoS metric, we show that local adaptive filters can deliver very significant performance gains relative to last-measure and moving-average predictors. We also show how global adaptive filters can produce performance gains, albeit at a lower level. These results show that adaptive prediction techniques have a significant role to play in the QoS Map construction.