Catálogo de publicaciones - libros

The introduction of the smart home has been seen as a way of allowing elderly people to lead an independent life for longer, making sure they remain safe and in touch with their social and care communities. The assistance could be in the form of helping with everyday tasks, e.g. notifying them when the milk in the fridge will be finished or institute safeguards to mitigate risks. In order to achieve this effectively we must know what the elderly person is doing at any given time. This paper describes a tiered approach to deal with recognition of activities that addresses the problem of missing sensor events that can occur while a task is being carried out.

In this paper, we propose a context framework reference model that aims at supporting distributed context aggregation, processing, provision and usage in ubiquitous computing environments. Based on this reference model, we introduce our design of GlobeCon, a scalable context management framework that can be embedded into context aware systems operating in local as well as wide area networks. One of the key aspects of GlobeCon is that it aggregates context information from largely distributed, heterogeneous and disparate sources in a hierarchical manner and presents context aware systems with a unified context access interface. GlobeCon is scalable in numerical, geographical and administrative dimensions, which is achieved by three means: (i) organizing context aggregation hierarchically, (ii) effectively distributing the load of context collection and context processing (i.e., interpreting and reasoning) among local context managers, and (iii) providing a dynamic federation of the distributed directory servers.

Supporting adaptive processes in tackling emergency situations, such as disasters, is a key issue for any emergency management system. In such situations, various teams are deployed in many sites, at the front-end or back-end of the situations, to conduct tasks handling the emergency. Context information is of paramount importance in assisting the adaptation of these tasks. However, there is a lack of middleware supporting the management of context information in emergency situations. This paper presents a novel framework that manages and provides various types of context information required for adapting processes in emergency management systems.

Context-aware applications require context information to adapt their behaviour to the current situation. When developing context-aware applications, application developers need to transform specific application context requirements into application logic to discover, select and bind to suitable sources of context information. To facilitate the development of context-aware applications, we propose a Context Binding Transparency that simplifies the process of retrieving context information. A major element of this transparency is the declarative approach to capturing context requirements. This enables application developers to specify their context requirements at a high level of abstraction rather than in programming code, and thus to separate the transformation of context requirements into context binding logic from the development of the actual application logic. In this way, we try to decrease the development effort and facilitate maintenance and evolution of context-aware applications. This paper discusses the design of this binding transparency; especially focusing on the language we developed to capture context requirements.

Despite progress in the development of context aware applications and supporting systems, there is still significant diversity in the models and abstractions they expose. This work describes an effort to gain a better understanding of the situation and develop a core set of abstractions by deploying several context aware applications, using a rapid prototyping platform. From this experience we propose and demonstrate a set of abstractions shown to be useful for a range of context aware applications. Combined with a survey and analysis reported elsewhere [1] we then provide an analysis toward providing a core set of abstractions that we argue can be used as the basis for modeling many context aware systems, including not only context, but other aspects such as entities, their relationships and associated events, services and content. We then provide several practical lessons learned from the use of our model and abstractions during analysis and our iterative platform development process.

The recent introduction of passive RFID tags into leading retailers’ supply chains has had a tremendous impact on their ability to manage the flow of goods. Tags on cases and pallets increase the supply chain’s visibility and allow for accurate tracking and tracing. Currently, additional efforts are being made to use RFID to actively monitor a product’s shipping condition (e.g. temperature, shock-vibration, etc.). This is of substantial interest for perishable goods and pharmaceuticals, whose shipping conditions are regulated. For continuous tracking of environmental data during the process of the movement of goods through the supply chain, RFID tags need to be equipped with sensors. Such sensors require continuous power, which is usually supplied by a battery on the tag. However, a battery not only significantly increases the cost of the hardware and makes it heavy and bulky, but also limits a tag’s lifetime. We propose ambient energy scavenging as a method to power sensors on battery-free RFID tags for continuous temperature monitoring and we show its applicability to the cold chain. Through detailed analysis of typical transport conditions we have identified ambient power sources which allow us to specify chip requirements and to make informed decisions about tag placement and total cost. We conclude that efficient monitoring ability is available at significantly lower cost than comparable implementations with active tags. Due to reduced costs, we predict high market penetration, which will result in more detailed information about multi-echelon supply chains. The fine-grained measurements will reveal failures and inefficiencies in the cold chain at a level of detail that would be hard to achieve with active tags. The elimination of the short-comings in the cold chain will result in reduced shrinkage, better quality and freshness of goods, and an overall reduction of losses of revenue.

We present a new approach for the detection of complex events in Wireless Sensor Networks. Complex events are sets of data points that correspond to interesting or unusual patterns in the underlying phenomenon that the network monitors. Our approach is inspired from time-series data mining techniques and transforms a stream of real-valued sensor readings into a symbolic representation. Complex event detection is then performed using distance metrics, allowing us to detect events that are difficult or even impossible to describe using traditional declarative SQL-like languages and thresholds. We have tested our approach with four distinct data sets and the experimental results were encouraging in all cases. We have implemented our approach for the TinyOS and Contiki Operating Systems, for the Sky mote platform.

We are developing a sensor network to assess the hydro-dynamics of surface water drainage into Great Crowden Brook in the Peak District (UK). The complete network will observe soil moisture, temperature and rainfall on a number of vertical slope transects. GSM access for remote real time reporting of network status is only available from the hilltops so a multihop communication strategy is being used. To minimise radio usage and maximise battery life we are reporting only those alarms and events that are judged to be of high priority by a simple rule based decision engine, based on spatio-temporal cross-correlation of the available sensor inputs. In this paper we present the data handling strategy, report the findings from the initial technology trial and discuss the implications of the recovered environmental data samples for the design of effective alarm generation rules. It is clear that the measurements can always be interpreted more reliably when richer contextual information is captured, but care must be taken with the choice of observables.