Rico Kusber

Measuring the Probability of Correctness of Contextual Information in Context Aware Systems

Context-awareness refers to computing systems that are able to sense and to comprehend their environment in order to adapt themselves in dependence to the available and relevant contextual information they depend upon. Gathering such contextual information involves real world entities such as sensors, which, for various reasons, are often prone to certain degrees of uncertainty and inaccuracy. Nevertheless, high quality context information plays a vital role in ensuring correct system behavior as well as dynamic system and service adaptation. Thus, a set of indicators is required which allows determining the quality of contextual information, which is commonly known as Quality of Context (QoC). One of the most relevant parameter of the QoC is the Probability of Correctness (PoC), which expresses the level of confidence, that the contextual information sensed, are in fact correct or not. In this paper, we propose an approach for measuring the PoC of context information by firstly analyzing the nature of context information and, secondly, revisiting the concept of Quality of Context also discussing other QoC parameters. Finally, we present a novel approach for quantifying the PoC for specific context information and evaluate the proposed method on a concrete case study.

Self-Organized Data Ecologies for Pervasive Situation-Aware Services: The Knowledge Networks Approach

Pervasive computing services exploit information about the physical world both to adapt their own behavior in a context-aware way and to deliver to users enhanced means of interaction with their surrounding environment. The technology to acquire digital information about the physical world is becoming more available, making services at risk of being overwhelmed by such growing amounts of data. This calls for novel approaches to represent and automatically organize, aggregate, and prune such data before delivering them to services. In particular, individual data items should form a sort of self-organized ecology in which, by linking and combining with each other into sorts of “knowledge networks” (KNs), they are able to provide compact and easy-to-be-managed higher level knowledge about situations occurring in the environment. In this context, the contribution of this paper is twofold. First, with the help of a simple case study, we motivate the need to evolve from models of “context awareness” toward models of “situation awareness” via proper self-organized “KN” tools, and we introduce a general reference architecture for KNs. Second, we describe the design and implementation of a KN toolkit that we have developed, and we exemplify and evaluate algorithms for knowledge self-organization integrated within it. Open issues and future research directions are also discussed.

DCCLA: Automatic Indoor Localization Using Unsupervised Wi-Fi Fingerprinting

People spend most of their time in a few significant places and often indoors in a small number of select rooms and locations. Indoor localization in terms of a users current place, related to a users daily life, routines or activities, is an important context. We implemented an automatic approach DCCLA Density-based Clustering Combined Localization Algorithm to automatically learn the Wi-Fi fingerprints of the significant places based on density-based clustering. In order to accommodate the influence of the signal variation, clustering procedure separately works on a list of RSSIs Received Signal Strength Indicators from each AP Access Point. In this paper, the approach is experimentally investigated in a laboratory setup and a real-world scenario in an office area with adjacent rooms, which is a key challenge to distinguish for place learning and recognition approaches. From these experiments, we compare and identify the most suitable parameters for the unsupervised learning.

An Approach to Autonomic Deployment Decision Making

Adding autonomicity to computing systems seems to be a promising way to deal with the problem of increasing system complexity. One step along the way to self-managing computing systems --- especially with regard to distributed, modularized, service based environments --- is to solve the problem of how to autonomically decide in a most useful and resource efficient way which alternative to choose in order to deploy a service. Deploying a service means, to either copy or move it from a source to a destination device or to use it remotely. In this paper we motivate the domain of autonomic service deployment and present an approach for deployment decision making (DDM). We explain all steps of the deployment decision making process and assemble them into an algorithm accordingly. Furthermore, we define all necessary components of DDM and enumerate a set of research questions which we address in order to fully explore the concerned domain. An experiment illustrates the potential of the presented approach.

Adaptive Services in a Distributed Environment

The dynamically changing nature of distributed service environments requires a computing model in which services modify their behavior in order to adapt to changes and fulfill new requirements in the environment. This paper describes design and implementation of the autonomic communication element (ACE) component model with main focus on service adaptability. Autonomic Communication Element (ACE) is a component model developed to support the creation of autonomic communication services which are capable of adapting their behavior according to changes in the context in which they are executing and from which they are requested. In this paper we present our approach for service adaptability and illustrate this on the example of an ACE.

Direction Detection of Users Independent of Smartphone Orientations

Smartphone sensors deliver useful information for applications such as indoor and outdoor navigation. An integral part of such applications is the detection of the orientation and movement direction of a smartphone user. Until now, movement direction detection using smartphones typically relies on GPS, which is often not available indoors. Alternatively, other approaches use sensors such as accelerometer and compass instead. These approaches rely on carrying the smartphone in a predefined orientation, or knowing the orientation of the smartphone in relation to the orientation of the user. In this paper, we present an approach to detecting the orientation and movement direction of users carrying smartphones inside the trouser pocket. This approach first determines the orientation of the smartphones top using compass and orientation sensor. Second, this approach determines the orientation of the smartphones screen, and the users movement direction by observing compass and accelerometer during at least two steps the user takes. After these two steps, the approach is capable of continuously aligning smartphone orientation and user orientation. With our approach, the user is free to change direction, movement speed, or to stop moving at all. The smartphone can be placed in the trouser pocket arbitrarily. And the smartphone is free to wobble in the trouser pocket. How well our approach works, is investigated based on experimental measurements.

3 rd workshop on recent advances in behavior prediction and pro-active pervasive computing

The 2nd Workshop on recent advances in behavior prediction and pro-active pervasive computing focuses on contributions that target recent challenges of context prediction and on applications of context prediction. The main challenges are a lack of benchmarks and common data sets, as well as a lack of development frameworks and that the main focus of context prediction still remains location prediction. Since context prediction is a key requirement to enable proactive applications, the workshop aims to intensify the discussion about the state and direction of context prediction research and to facilitate collaboration among research groups focusing on context prediction.

Self-adaptation for Deployment Decision Making

In this paper we present an intelligent assistant for deploying services and obtaining content in distributed computing network. We focus on how self-adaptation is achieved and illustrate the capabilities and limits of self-adaptation within our approach with the help of provider reputation values.

Implicit positioning using compass sensor data

In this paper, we present Implicit Positioning -- an approach to recognize the indoor area where a person is, such as a specific corridor or corner. Implicit Positioning adapts to its user by automatically finding, learning, and recognizing patterns in data from a smartphones compass sensor. These patterns model characteristics of the areas the person walked through, such as the direction the person headed, and can be understood as Implicit Positions. Implicit Positioning neither relies on user feedback nor on additional infrastructure, digital maps or war-driving. All the person needs to do is carry a smartphone. To evaluate our approach, we collected data from smartphones carried by a test person in different positions. The test person covered a distance of 17.4 km in two different buildings. When taking a path for the second time, patterns were already being correctly recognized with an accuracy of up to 87.8%.

Feature weighting for CPM-based context validation

Context Pattern Method (CPM) is a statistical method that is used to quantify the validity of contextual information based on dependent contexts using previous knowledge about the system. The method exploits the interdependencies in a context aware system among entities and the environment in which they reside in order to calculate the Probability of Correctness (PoC) for a context under investigation. PoC expresses the level of confidence, that the contextual information sensed, are in fact correct or not. Obviously, each of the dependent contexts has a different importance to the context that is under investigation. Therefore its influence to the PoC measure needs to be weighted accordingly. In this paper we discuss the concept of feature weighting and show how feature selection algorithms can be applied for this purpose. We apply chi2, relief-f and mutual information, algorithms to the CPM method in order to weight the influence of the individual dependent contexts to the overall PoC measure and evaluate the methods performance.