Ebook: Workshop Proceedings of the 8th International Conference on Intelligent Environments

For the fourth time, the Intelligent Environments Conference hosted a range of international and multidisciplinary workshops. The conference itself was held for the eighth time in 2012 at the historic World Heritage UNESCO city of Guanajuato (Mexico), following a series of highly successful conferences that were organized in Colchester, UK (2005), Athens, Greece (2006), Ulm, Germany (2007), Seattle, USA (2008), Barcelona, Spain (2009), Kuala Lumpur, Malaysia (2010), and Nottingham Trent University, United Kingdom (2011). The workshops provide a forum for scientists, researchers and engineers from both industry and academia to engage in discussion on newly emerging, or rapidly evolving topics that have yet to reach the level of maturity associated with conferences or journals. This year we are pleased to include the following workshops:

WISHWell'12: The 4th International Workshop on Intelligent Environments Supporting Healthcare and Well-being discussed novel research regarding all aspects of pervasive healthcare solutions. The workshop was organized by Juan Carlos Augusto (University of Ulster, United Kingdom) and John O'Donoghue (University College Cork, Republic of Ireland).

AITAmI'12: The 7th workshop on Artificial Intelligence Techniques for Ambient Intelligence aimed at gathering researchers in a variety of AI subfields together with representatives of commercial interests to explore the technology and applications for ambient intelligence. This year's edition presented a special focus on context and situation understanding. The workshop was organized by Asier Aztiria (University of Mondragon, Spain), Juan Carlos Augusto (University of Ulster, United Kingdom), and Diane Cook (Washington State University, USA).

WOLSIE'12: The Workshop on Large Scale Intelligent Environments was organized by James P. Dooley (Essex University), Matthew H. Ball (Essex University), and Mohammad Al-Mulla (Kuwait University). The aim of the workshop was to go towards intelligent environments beyond the size of normal experimental labs.

iDR'12: The First International Workshop on Intelligent Domestic Robots provided a forum to explore the latest research in domestic robots, with particular emphasis on robots that are functionally designed to interact and communicate with people. The workshop was organized by Simon Egerton (Monash University, Sunway Campus, Malaysia) and Eduardo Morales (National Institute for Astrophysics, Optics and Electronics, Mexico).

WOFIEE'12: The 1st Workshop on Future Intelligent Educational Environments was organized by Victor Callaghan (Essex University), Minjuan Wang (San Diego State University), and Juan Carlos Augusto (University of Ulster, United Kingdom). The focus on the workshop is on the increasingly crucial role intelligent environment technology is playing in the delivery of education.

IMIASH'12: The 1st International Workshop on Intelligent Multimodal Interfaces Applied in Skills Transfer, Healthcare and Rehabilitation investigated the strong integration between humans and technology as an opportunity to investigate human behavior through this interaction and to analyze potentialities and advantages when the multimodal interfaces are used in the field of skills transfer, healthcare and rehabilitation. The organizers were Oscar Sandoval (Instituto Tecnológico de Orizaba, Mexico), Alejandro Rodriguez-González (University Carlos III of Madrid, Spain), Paolo Tripicchio (Perceptual Robotics Labortory (PERCRO) Scuola Superiore Sant Anna, Italy), and Otniel Portillo-Rodríguez (University of Mexico State, Mexico).

WoRIE'12: The 2nd Int. Workshop on Reliability of Intelligent Environments was organized by Miguel J. Hornos (University of Granada, Spain), Juan Carlos Augusto (University of Ulster, United Kingdom), and Pablo A. Haya Coll (Escuela Politécnica Superior IIC-UAM, Spain). The event brought together developers and researchers to focus on all aspects of the development process that can contribute to make Intelligent Environment systems safer and to provide methodologies that can increase the confidence in these developments.

AIE'12: The 1st International Workshop on Improving Industrial Automation Using the Intelligent Environments Paradigm was organized by Juan Carlos Orozco (ACELAB Industrial Automation, México) and Victor Zamudio (Instituto Tecnológico de León, México). It addressed the oportunities brought by the intelligent environments paradigm to industrial automation. The format of this workshop was different than the rest. It had no paper based presentations in order to foster discussion among participants coming from the industry rather than from an academic context.

As is evident from this list, the IE'12 workshops addressed a diverse set of topics. The formats are equally diverse ranging from regular lectures to practical sessions.

Of course, we would like to take the opportunity to thank everybody who made these proceedings possible. We are especially grateful to our organizing committees, without whom these workshops would not have been possible. We are also grateful to the local organizers who worked tirelessly behind the scenes to make these events a success. Of course we reserve our special thanks for the researchers, who do the work, achieve the advances, set the research agenda and then come to Intelligent Environments workshops to present and discuss their insights. Through these proceedings we hope to be able to share not only these insights but also the unique, inspiring spirit of the IE workshops.

Juan A. Botía, Hedda Schmidtke, Tatsuo Nakashima

General chairs of the IE'12 Workshops

Pervasive healthcare aims at applying mobile and ubiquitous computing technologies to address some of the major issues facing healthcare, such as the aging of the population, an increase in the prevalence of chronic diseases and escalating costs. One of the main characteristics of pervasive healthcare systems is their reliance on contextual information such as using the location of a nurse in a hospital ward to provide her with information relevant to the task at hand, or reminding a patient to take his medication when he returns home and his pillbox has not been opened during the day. Advances in activity recognition have made possible a second generation of pervasive healthcare solutions using activity-aware computing to identify the activities of daily living being performed by an elder and assist her with the task, or to monitor the intensity of the physical activity performed by a patient undergoing treatment. Moving forward this paper describes a new wave of pervasive healthcare applications, which are becoming possible due to progress in the estimation of user behavior. These behavior-aware systems can be used to measure the effects of medication in patients suffering from mental disorders, provide feedback in rehabilitation therapies or to perceive abnormal behaviors that provide early signs of an ailments such as Alzheimer's disease.

We report the development of an architecture for a Multi-Agent Social Sound which is based on the human personality in the search for empathy among persons focused on mobile robotics for evaluation of working groups in different controlled environments, focusing specifically on how they affect emotions to personality in a situation before making the stimulus act (emotion), taking on the work in the psychology of personality which Raymond Cattell [13] developed a tool called the 16 PF for the study of personality in a short time, emotions using the methodology of human primary emotions of Paul Ekman [12] who lets us know that there are six primary emotions Joy, Surprise, Anger, Fear, disgust and sadness, the paradigm of Intelligent Agents.

In this article we present a comparative study between a Bayesian classifier and the method SPC (Statistical Process Control). The Bayesian classifier is included in the virtual machine WEKA, and SPC is manly used to monitor processes in different domains. These two methods are applied to a set of data collected from a SunSPOT in order to detect falls. A comparative analysis of these two approaches was performed in order to have the best strategy to be apply in real scenarios.

In this study a partitional clustering technique is proposed. Our proposal is a variant of the c-Means algorithm that replaces its traditional minimum-distance classifier by a classifier based on associative memories. The variant was compared against the original version by applying both techniques to three datasets belonging to the UCI Machine Learning Repository: Iris, Wine and Pima Indian Diabetes. As a comparison criterion, an intracluster-spread index was used. Results obtained in experimental tests show that, when applied to certain databases, the PAC-Means technique overcomes to the c-Means algorithm.

The system presented in this paper aims to serve as a low cost tool for detection and characterization of different neurodegenerative diseases by measuring the skeletal muscle movement in the patients´ fingers. Conditions like Parkinson's, Multiple Sclerosis, psychiatric diseases, alcoholism and drug abuse have specific movement frequencies, which could be graphed. The system looks to make possible the identification of the disease mapped according to the characteristic graph, reducing the number of tests needed for this matter. Research focused, it aims to measure the disease's intensity and progress as well as the effect of medicine treatment and therapy.

The purpose of this paper is to present a method that utilizes lung sounds (LS) for quantitative assessment of patient health, based on the fact that LS are relates to respiratory disorders. Traditional asthma evaluation methods may involve auscultation and spirometry. However, improved diagnosis opportunities can be offered by utilizing sensitive electronic stethoscopes (now widely available), and the application of quantitative signal analysis methods. In this context, we carried out experiments using normal LS from both the RALE repository and recordings from students. In this paper, we propose an acoustic evaluation methodology based on the Gaussian Mixed Models (GMM) that should assist in broader analysis, identification, and differentiation of LS. Additionally, frequency domain analysis of peculiar sounds reflected during wheezes and crackles, and their differentiation from normal respiratory sounds should assist in improved diagnosis.

As part of our efforts to improve the effective design of suitable CS applications for the elderly, during the last three years we have been performing an ensemble of studies, which have enabled us to understand the nature of CS activities and to envision how we might support them by means of Ambient Intelligence (AmI) technologies. Furthermore, we have proposed UCSA, a framework for the design of usable CS applications for worried-well elders. In order to validate the framework, we have developed and evaluated the usability of several prototypes by means of TAM-based questionnaires. In this work, we present the results of a further evaluation of one such application, which focuses on user experience (UX). We identified which kinds of interactions occurred among the different participants, along with their form and functions, as a means to gather evidence about the application's user experience. Finally, we discuss the implications of these findings regarding the framework's UX perspective, along with possible directions of future work.

Adding reasoning abilities to context-aware systems has been a focus of research in pervasive computing for several years and a broad range of approaches has been suggested. In particular, the well-known trade-off between expressivity and inferential power has been discussed as a major concern, as dimensions of context include well-known hard domains, such as spatial, temporal, and causal knowledge. In practice however, experiments report acceptable run-times and complexity of the used reasoning mechanism does not seem to be an issue at all. Two questions are addressed in this paper: why this is the case and whether these positive results will scale up as context-aware systems are leaving their experimental environments, are extended and modified by application developers, and employed in everyday life by millions of end-users.

The paper presents an analysis of results from pervasive computing, qualitative spatial and temporal reasoning, and logic-based contextual reasoning. The goal was to carve out the theoretical core of fast contextual reasoning reported from experimental context-aware systems, and to discuss how these good properties can be made to scale up. The findings suggest that partial order reasoning is the core of tractable contextual reasoning. Examples illustrate the surprisingly high expressiveness and inferential power, and serve to emphasize the interrelations between tractable reasoning in pervasive computing, qualitative spatial and temporal reasoning, and logic-based contextual reasoning.

In this paper, we present an implementation of a real time activity recognizer running on a cellphone. First, simple activity recognition from accelerometer data is performed and then, this information is fused with data from Wi-Fi Access Points to classify the activity being performed by the user. The training set consisted of 8 activities performed in an academic environment and the classification accuracy was 89.7% using a supervised learning approach.

This paper presents a methodology for classification of 3D objects that can be useful to solve some problems in context-awareness, automatic surveillance and ambient assisted living.

This methodology uses light line images projected onto the object being studied. The use of structured light allows to get the three-dimensional object information. The images, that were captured by a CCD camera, were improved with tools of digital image processing and were characterized by three compactness measures. The classification was performed by some algorithms like Naive Bayes, Neural Nets and an associative classifier.

This research is aimed to get an intelligent navigation robot, whose functionality is based on the interaction of two agents, a reactive agent and a deliberative agent, where both agents take decisions jointly to achieve a goal. The deliberative agent makes a decision based on the acquired knowledge and determines which way leads to the best result. The reactive agent receives the decision made by the deliberative agent and transforms it into a physical interaction with the environment using line following techniques. This method was implemented successfully into a robot and the tests results were as expected.

Given a set of agents with valid previous knowledge bases obtained by a generator of beliefs, we wish to know how new knowledge affects each agent. To model the new knowledge, boolean logic is used, expressed by 2CNF clauses, to reduce the complexity. Upon recieving new knowledge, one or more agents may find it inconsistent with their previous knowledge base, so a mechanism is applied which removes knowledge by using a contraction operation,described by the AGM model in order to ensure consistency of the knowledge base. The goal is to determinate if that contradicting knowledge significantly affects the set of beliefs of each agent.

Furthermore, a problem is modeled in which, given a set of agents (production operators) and their knowledge base (skills in the industry), clauses represent new activities are added when a new process is required and the model determines which is the most suitable agent to perform the activities, through from an evaluation mechanism of the inconsistencies that was generated.

We have developed software which allows to support make decisions over the more capable workers to develop an activity based on their previous skills (knowledge).

Cyclic instability is a problem that, despite being shown to affect intelligent environments and in general any rule-based system, the strategies available to prevent it are still limited and mostly focused on centralized approaches. These approaches are based on topological properties of the Interaction Network (IN) associated, and locking a set of agents. In this paper we present a comparative study of the performance of different optimization techniques when solving the problem of cyclic instability in synthetic scenarios. Instead of using the Interaction Network of the System (which can be computationally expensive, specially in very dense systems). We introduced the concept of Average Change of the System (ACS) in order to measure the oscillatory behavior of the system and an optimization strategy. In particular Particle Swarm Optimization (PSO), Micro-Particle Swarm Optimization (μ-PSO), Bee Swarm Optimization (BSO), Artificial Immune Systems (AIS) and Genetic Algorithms (GA) were considered. The results found are very promising, as they can successfully prevent unwanted oscillations. Additionally, some of these strategies could be implemented using parallel and distributed processing, in order to be used in real-time scenarios.

This article develops a method of fuzzy fingerprint pre-processing to apply in a pattern recognition system, in which a neural network is developed with fuzzy response integration for fingerprint recognition. There are also shown neural network training results using Biometrics Laboratory database at the Bologna University and of using a genetic algorithm, as well as the chromosome used in the genetic algorithm development of membership functions optimization of fingerprint pre-processing fuzzy method.

As Intelligent Environments research continues in its various guises, there is an increase in the scale at which investigation is being conducted. Projects are looking beyond single-room and apartment sized living-labs that have provided us with initial testbeds for early research and are envisioning grander designs on the scale of entire buildings, campus or towns. But now we must ask if the knowledge we have gained previously can scale upwards, or are new methods and models required to break free from the confines of controlled labs and into real-world deployments where multi-user is the norm not the exception. In the discourse of this paper, we describe what Large-Scale Intelligent Environments (LSIEs) are and identify some of the challenges that relate to realizing them. We highlight the importance of security, standards and infrastructure so that human users can roam confidently from place-to-place whilst enjoying a seamless continuity of experience.

Smart home technologies hold promise for many aspects of daily life. Research and development of these systems has matured for elder care, energy efficiency, and home safety applications. The focus of most implementations has been on single living spaces for a small number of individuals. New low-power wireless systems, inexpensive computing power, and widely available network access has reached the point where large scale ubiquitous computing technologies have become much more feasible.

The smart environments research community has few projects to explore issues and techniques for deploying large scale ubiquitous systems. This work summarizes some of the existing works and introduces the Smart Home in a Box (SHiB) Project. The upcoming SHiB Project targets building 100 smart homes in several kinds of living spaces for gathering longitudinal data from a significant number of residents. The resulting data set will provide opportunities to answer open questions in the areas of transfer learning, active learning, digital asset migration, middleware architectures, activity detection and discovery, human factors, smart home installation, and others.

We envision an Ambient Intelligent Environment as an environment with technology embedded within the framework of that environment to help enhance an users experience in that environment. Existing implementations , while working effectively, are themselves an expensive and time consuming investment. Applying the same expertise to an environment on a monolithic scale is very inefficient, and thus, will require a different approach. In this paper, we present this problem, propose theoretical solutions that would solve this problem, with the guise of experimentally verifying and comparing these approaches, as well as a formal method to model the entire scenario.

The trend towards implementing applications supporting the “Internet of Things” (IoT) concept is allowing the discovery of amazing new opportunities that current technologies could bring to us. Previous efforts have shown the challenges yet to be solved. Research for increasing the capabilities of the IoT is being conducted under various paradigms and analyzing different approaches for finding useful ways to implement valuable applications under the IoT concept. This paper proposes a framework based on a MultiAgent Systems Paradigm, which is an accepted form for modeling and implementing applications that are strongly related real-world concepts. The idea behind the proposal is to simplify the development of applications in the IoT which could perform complex tasks in a distributed environment.