Graeme Stevenson

Self-aware Pervasive Service Ecosystems

Here we present the overall objectives and approach of the SAPERE (“Self-aware Pervasive Service Ecosystems”) project, focussed on the development of a highly-innovative nature-inspired framework, suited for the decentralized deployment, execution, and management, of self-aware and adaptive pervasive services in future network scenarios.

A top-level ontology for smart environments

Recognising human activities is a problem characteristic of a wider class of systems in which algorithms interpret multi-modal sensor data to extract semantically meaningful classifications. Machine learning techniques have demonstrated progress, but the lack of underlying formal semantics impedes the potential for sharing and reusing classifications across systems. We present a top-level ontology model that facilitates the capture of domain knowledge. This model serves as a conceptual backbone when designing ontologies, linking the meaning implicit in elementary information to higher-level information that is of interest to applications. In this way it provides the common semantics for information at different levels of granularity that supports the communication, reuse and sharing of ontologies between systems.

Developing pervasive multi-agent systems with nature-inspired coordination

Pervasive computing systems can be modelled effectively as populations of interacting autonomous components. The key challenge to realizing such models is in getting separately-specified and -developed sub-systems to discover and interoperate with each other in an open and extensible way, supported by appropriate middleware services. In this paper, we argue that nature-inspired coordination models offer a promising way of addressing this challenge. We first frame the various dimensions along which nature-inspired coordination models can be defined, and survey the most relevant proposals in the area. We describe the nature-inspired coordination model developed within the SAPERE project as a synthesis of existing approaches, and show how it can effectively support the multifold requirements of modern and emerging pervasive services. We conclude by identifying what we think are the open research challenges in this area, and identify some research directions that we believe are promising.

Ontonym: a collection of upper ontologies for developing pervasive systems

Pervasive systems present the need to interpret large quantities of data from many sources. Context models support developers working with such data by providing a shared representation of the environment on which to base this interpretation. This paper presents a set of requirements for a context model that addresses uncertainty, provenance, sensing and temporal properties of context. Based on these requirements, we describe Ontonym, a set of ontologies that represent core concepts in pervasive computing. We propose a framework for evaluating ontologies in the pervasive computing domain by combining recognised techniques from the literature, and present a preliminary evaluation of Ontonym using these criteria.

KCAR : a knowledge-driven approach for concurrent activity recognition

Abstract Recognising human activities from sensors embedded in an environment or worn on bodies is an important and challenging research topic in pervasive computing. Existing work on activity recognition is mainly concerned with identifying single user sequential activities from well-scripted or pre-segmented sequences of sensor events. However a real-world environment often contains multiple users, with each performing activities simultaneously, in their own way and with no explicit instructions to follow. Recognising multi-user concurrent activities is challenging, but essential for designing applications for real environments. This paper presents a novel Knowledge-driven approach for Concurrent Activity Recognition (KCAR). Within KCAR, we explore the semantics underlying each sensor event and use semantic dissimilarity to segment a continuous sensor sequence into fragments, each of which corresponds to one ongoing activity. We exploit the Pyramid Match Kernel, with a strength in approximate matching on hierarchical concepts, to recognise activities of varying grained constraints from a potentially noisy sensor sequence. We conduct an empirical evaluation on a large-scale real-world data set that is collected over one year and consists of 2.8 millions of sensor events. Our results demonstrate that KCAR achieves an average recognition accuracy of 91%.

LOC8: A Location Model and Extensible Framework for Programming with Location

Location is a core concept in most pervasive systems-and one thats surprisingly hard to deal with flexibly. Using a location model supporting a range of expressive representations for spaces, spatial relationships, and positioning systems, the authors constructed LOC8, a programming framework for exploring location datas multifaceted representations and uses. With LOC8, developers can construct complex queries by combining basic queries and additional contextual information.

Semantic web technologies in pervasive computing

Pervasive and sensor-driven systems are by nature open and extensible, both in terms of input and tasks they are required to perform. Data streams coming from sensors are inherently noisy, imprecise and inaccurate, with differing sampling rates and complex correlations with each other. These characteristics pose a significant challenge for traditional approaches to storing, representing, exchanging, manipulating and programming with sensor data. Semantic Web technologies provide a uniform framework for capturing these properties. Offering powerful representation facilities and reasoning techniques, these technologies are rapidly gaining attention towards facing a range of issues such as data and knowledge modelling, querying, reasoning, service discovery, privacy and provenance. This article reviews the application of the Semantic Web to pervasive and sensor-driven systems with a focus on information modelling and reasoning along with streaming data and uncertainty handling. The strengths and weaknesses of current and projected approaches are analysed and a roadmap is derived for using the Semantic Web as a platform, on which open, standard-based, pervasive, adaptive and sensor-driven systems can be deployed.

USMART: An Unsupervised Semantic Mining Activity Recognition Technique

Recognising high-level human activities from low-level sensor data is a crucial driver for pervasive systems that wish to provide seamless and distraction-free support for users engaged in normal activities. Research in this area has grown alongside advances in sensing and communications, and experiments have yielded sensor traces coupled with ground truth annotations about the underlying environmental conditions and user actions. Traditional machine learning has had some success in recognising human activities; but the need for large volumes of annotated data and the danger of overfitting to specific conditions represent challenges in connection with the building of models applicable to a wide range of users, activities, and environments. We present USMART, a novel unsupervised technique that combines data- and knowledge-driven techniques. USMART uses a general ontology model to represent domain knowledge that can be reused across different environments and users, and we augment a range of learning techniques with ontological semantics to facilitate the unsupervised discovery of patterns in how each user performs daily activities. We evaluate our approach against four real-world third-party datasets featuring different user populations and sensor configurations, and we find that USMART achieves up to 97.5p accuracy in recognising daily activities.

Sapphire: Generating Java Runtime Artefacts from OWL Ontologies

The OWL ontology language is proving increasingly popular as a means of crafting formal, semantically-rich, models of information systems. One application of such models is the direct translation of a conceptual model to a set of executable artefacts. Current tool support for such translations lacks maturity and exhibits several limitations including a lack of support for reification, the open-world assumption, and dynamic classification of individuals as supported by OWL semantics. Building upon the state-of-the-art we present a mapping from OWL to Java that addresses these limitations, and its realisation in the form of a tool, Sapphire. We describe Sapphire’s design and present a preliminary evaluation that illustrates how Sapphire supports the developer in writing concise, type safe code compared to standard approaches while maintaining competitive runtime performance with standard APIs.

Towards a reliable, wide-area infrastructure for context-based self-management of communications

In this paper we describe ConStruct, a distributed, context-aggregation based service infrastructure which supports the development of context-aware applications. ConStruct operates by automatically generating and maintaining directed context-processing graphs which connect applications to the sources of data they require at a relevant level of abstraction. The infrastructure also supports the dynamic creation of context processing elements to bridge gaps between available and requested information. ConStruct provides a reliable, scalable infrastructure; focused on self-maintenance in order to alleviate developer workload. We describe the infrastructure design and implementation, the associated programming model, and our planned extensions to the infrastructure.