Ambient Intelligence

This paper is aimed to introduce IDIAP Brain Computer Interface (IBCI) research that successfully applied Ambience Intelligence (AmI) principles in designing intelligent brain-machine interactions. We proceed through IBCI applications describing how machines can decode and react to the human mental commands, cognitive and emotive states. We show how effective human-machine interaction for brain computer interfacing (BCI) can be achieved through, 1) asynchronous and spontaneous BCI, 2) shared control between the human and machine, 3) online learning and 4) the use of cognitive state recognition. Identifying common principles in BCI research and ambiance intelligence (AmI) research, we discuss IBCI applications. With the current studies on recognition of human cognitive states, we argue for the possibility of designing empathic environments or devices that have a better human like understanding directly from brain signals.

The vision behind the Personalized Living Environment (PLE) is that of recreating a user's customized living or working surroundings in accordance with user-defined settings and in response to the user's mobility. Concepts such Ambient Intelligence, Home Network, and Personal Environment Service (PES) could be instrumental in realization of PLE.. Indeed, integration of elements of these three technologies may be the catalyst for deployment of PLE and an enabler in attracting industrial attention. Moreover, underlying technologies related to the above concepts are envisioned to play a key role in the implementation of PLE, including mobile communications supported by ubiquitous infrastructure and Wireless Personal Area Networks. In this paper, we present a simple PES network architecture and its underlying network technologies, and we discuss a set of potential services of PES. We also compare PES with other related concepts capable of supporting the PLE vision.

Abstract. The Internet of Things allows people and objects to seamlessly interact, crossing the bridge between real and virtual worlds. Newly created spaces are heterogeneous; social relations naturally extend to smart objects. Conviviality has recently been introduced as a social science concept for ambient intelligent systems to highlight soft qualitative requirements like user friendliness of systems. Roughly, more opportunities to work with other people increase the conviviality. In this paper, we first propose the conviviality ...

We address some of the current challenges in intelli- gent interfaces for universal access and present the de- sign of an intelligent interface which is aimed at 1) input processing the user's sensory modalities (or modes) via various media, 2) building (or encoding) a model of user's emotions (MOUE) and 3) adapting its multime- dia output to provide the user with an easier and more natural technology access and interaction. We identify key research issues relating to one particularly rich ap- phcation of MOUE, namely home health care provided via telemedieine.

Signals coming out of human brain is considered as a new biometric. After acquisition, feature
extraction and classification of signals are done. To achieve this, spatial data/sensor fusion is done. EEG
features are classified using the neural network named as Elman neural network. 1600 raw EEG signals are
collected using maximum 35 informative channels and maximum recognition rate achieved is 98%. EEG
biometrics has got massive potential due to its sturdiness against any kind of scam.

Activity Related Biometrics based on motion trajectories Anastasios Drosou1, 2, Konstantinos Moustakas2, Dimos Ioannidis2, Dimitrios Tzovaras2 1Imperial College London, SW7 2AZ London, UK {a. drosou09@ imperial. ac. uk} 2CE. RTH-ITI, PO Box 361 57001 Thermi-...

In Ambient Intelligence (AmI) environments, some services provided by AmI devices are often not visible to users and to other devices. The existing approaches deal with services' composition and discovery as two independent parts. In this paper, we propose an alternative approach based on logical reasoning agent system. This system is supported by a communication protocol where agents discover automatically services provided in their environment and construct dynamically composite services. The service composition is constructed from an exchange of idiomatic expressions among agents and users, while the discovery process takes the form of an information request via the communication protocol. The advantage of this approach is that agents are able to acquire knowledge from each other and when interacting with users. This capability will facilitate the satisfaction of user's requirements in an intelligent way. This study shows that agents are able to satisfy new services previously unknown to the system.

Derailments, common in developing nations, often result from dislodged or defective rail blocks, leading to substantial property damage and loss of life. Developing an automated real-time wireless sensing system for preventing derailments is a complex challenge, particularly in resource-constrained regions with limited network infrastructure. Existing research has yet to provide a practical solution that effectively addresses the need for long-distance sensing and optimized sensor deployment. This research focuses on developing a comprehensive solution that addresses these challenges. We explore vibration sensing and multi-sensor fusion for accurate rail track detection. Through rigorous experimentation and analysis, we demonstrate a 95% accuracy rate in detecting incoming trains from a distance of 1 km. To optimize sensor deployment, we contrive a multi-objective optimization problem and employ a meta-heuristic approach. We generate effective sensor deployment topologies by considering factors such as cost, vulnerability, and accident statistics. Our findings reveal that optimized deployment can significantly reduce the deployment costs up to 50% while preventing over 65% of potential accidents compared to a brute-force deployment approach. This research offers a promising solution for real-time rail track detection, combining advanced technologies and data-driven optimization. The findings contribute to enhancing railway safety and mitigating the devastating consequences of derailments.

Wireless sensor networks (WSNs) have become ubiquitous, permeating every aspect of human life. In environmental monitoring applications (EMAs), WSNs are essential and provide a holistic view of the deployed environment. Physical sensor devices and actuators are connected across a network in environmental monitoring applications to sense vital environmental factors. EMAs bring together the intelligence and autonomy of autonomous systems to make intelligent decisions and communicate them using communication technologies. This paper discusses the various architectures developed for WSNs in environmental monitoring applications and the support for specific design goals, including machine learning in WSNs and its potential in environmental monitoring applications.

This paper presents a comprehensive study on the deployment strategies of Wireless Sensor Networks (WSNs) for enhancing spatial coverage and improving the accuracy and responsiveness of fire detection systems, particularly in large-scale and high-risk outdoor environments. Traditional fire detection approaches often suffer from limited coverage, delayed response times, and inadequate adaptability to varying terrains, rendering them less effective in rapidly evolving scenarios such as wildfires. To overcome these limitations, this research investigates and compares four key sensor deployment strategies: grid-based, cluster-based, random-based, and edge-based formations. Each strategy is evaluated in terms of its ability to achieve optimal sensor positioning, maximize network efficiency, and reduce blind spots in fire-prone regions. The experimental framework involves deploying these strategies within a clearly defined geographic area segmented into seven distinct regions. This structured approach enables systematic testing and simulation under controlled yet varied spatial conditions. Key performance metrics such as coverage efficiency, transmission frequency, responsiveness, energy consumption, and resource utilization are assessed to determine the relative strengths and weaknesses of each deployment model. Simulation results reveal that strategic deployment significantly enhances fire detection accuracy and improves network performance by minimizing detection delays and optimizing node distribution. Notably, the grid and clustered deployment approaches yield superior results in terms of balanced coverage and reduced redundancy, while edge-based configurations show potential in perimeter surveillance applications. The findings underscore the importance of adaptive deployment strategies in achieving scalable, energy-efficient, and resilient fire monitoring systems. In addition to providing comparative insights, the study aims to guide the design of next-generation WSN-based fire detection systems capable of adapting to complex environmental conditions. By aligning deployment strategies with application-specific needs, this research contributes to the realization of reliable, responsive, and cost-effective fire detection infrastructures that can significantly mitigate risks to public safety and property.

We present the Everyday Living Artificial Intelligence (AI) Hub, a novel proof-of-concept framework for enhancing human health and wellbeing via a combination of tailored wearable and Conversational Agent (CA) solutions for non-invasive monitoring of physiological signals, assessment of behaviors through unobtrusive wearable devices, and the provision of personalized interventions to reduce stress and anxiety. We utilize recent advancements and industry standards in Internet of Things (IoT) and AI technologies to develop this proof-ofconcept framework.

Ambient intelligent systems such as Home and Building Automation Systems (HBAS) are becoming evermore accepted and capable of actuating automatically on behalf of users to fulfil their requests or enable activities. However, when multiple users interact with such systems, the requirements of activities often interfere resulting in conflicting actuations which HBAS ought to automatically detect and resolve. Yet, despite recent advances in HBAS, no ambient intelligent solution has been reported that is adequately grounded on knowledge analysis. The contributions of this article are twofold. First, it reviews relevant literature on Ambient Intelligence, conflict detection, conflict resolution and knowledge representation in HBAS. Second, it proposes and validates an ontological framework for conflict detection and resolution backed by knowledge-based analysis. Effectively, the proposed solution performs automatic environment actuations maximizing users comfort and energy efficiency.

Home and Building Automation Systems (HBAS) are becoming of widespread adoption. When distinct users interact with such systems, their intentions are likely to be different, often resulting in conflicting situations, which the systems ought to recognize and resolve automatically. This work aims at investigating conflict in HBAS and creating a solution to detect and resolve them. Herein, we review the literature concerning conflict detection and resolution, and propose a formal framework based on constraint solving that enables detecting and solving conflict situations automatically.

Nous proposons, dans cet article, une méthode permettant de combiner des informations symboliques avec des informations numériques pour obtenir un processus d'estimation supervisé par un niveau d'analyse du contexte. L'application visée, ici, est la fusion de données. En effet les algorithmes implantés dans les systèmes multicapteurs doivent être conçus pour que le système fonctionne d'une façon nominale dans toutes les conditions opérationnelles. Pour cela le système doit adapter l'importance qu'il accorde à chaque capteurs, de sorte qu'à tout instant l'estimation tienne compte du contexte ; cette importance est accordée au vu de critères établis au préalable par un expert.

The XPod system, presented in this paper, aims to integrate awareness of human activity and musical preferences to produce an adaptive system that plays the contextually correct music. The XPod project introduces a "smart" music player that learns its user's preferences and activity, and tailors its music selections accordingly. We are using a BodyMedia device that has been shown to accurately measure a user's physiological state. The device is able to monitor a number of variables to determine its user's levels of activity, motion and physical state so that it may predict what music is appropriate at that point. The XPod user trains the player to understand what music is preferred and under what conditions. After training, the XPod, using various machine-learning techniques, is able to predict the desirability of a song, given the user's physical state.

Todays, we assist to the explosive development of mobile computing devices like PDAs and cell-phones, the integration of embedded intelligence (like Web server) in more and more common devices, and the proliferation of wireless communication technologies (IRdA, Bluetooth, IEEE 802.11, GPRS). All these trends contribute to move us closer to the ubiquitous computing world described by Mark Weiser. But while the technology is here, applications, and more important, models and tools for designing future ambient computing systems are still rare. One of the first innovative concept of ubiquitous computing, context-awareness, is still hard to use and understand from a programming perspective. We think that the problem resides in the lack of system support : in traditional computing, operating system offers simple to use and easy to understand abstractions of computational resources. Ubiquitous computing involves an integration of "computing" into the real-world, which is a radically different environment for applications. We think that this environment requires new operating system services and abstractions. Because the real world is made of physical entities, "living" in the physical space, ambient computing software should be able to use abstraction representing such objects, in a simple way. In this paper, we present a light framework to design ubiquitous computing software, called SPREAD. Unlike many approaches which hides too much of the real-world behind traditional computing abstraction, SPREAD defines programming abstraction based on the properties of the physical space. Hence, physical properties, like relative proximity, are used as implicitly in SPREAD as variable addressing in a computer memory. In SPREAD, application (or process) behavior can be "mechanically" driven, in the sense that actions flow can be directly dependent of physical mobility. To support this concept, we introduce a programming and execution model allowing to design computing and in-formation systems driven directly by arranging and moving physical objects in the space. We demonstrate the use of the model to implement a few practical applications, highlighting its simplicity and expression power.

There is growing evidence to suggest that the success of a co-evolutionary learning system may depend critically on the nature of the environment in which the learner is placed, and on certain attributes of the task domain, rather than the details of the particular learning algorithm employed. We discuss how a learning system can be modeled as a meta-level game between abstract entities which w e call performer, inltrator and evaluator. Learning can sometimes fail due to collusive suboptimal equilibria in this meta-game of learning. But some domains have special attributes which seem to prevent such collusions and thereby facilitate co-evolutionary advancement. A better understanding of these issues may help to improve the design of co-evolutionary learning systems in the future.

Web log data has been the basis for analyzing user query session behavior for a number of years, but it has several important shortcomings. The main one being that we dont really know what the user is doing -is s/he looking at the screen or doing something else? We have conducted an Eye-Tracking study to analyze user behavior when searching the web and looking for specific information on results and content pages. The goal is to obtain more precise information about the search strategy of the user. Which characteristics make the difference between successful and unsuccessful searches? This research presents results focusing on the number of formulated queries by session, documents clicked, the fixation durations on the documents, and the distribution of the attention in the different areas of the screen, among other aspects.

To my parents for providing endless energy. To my brothers for giving me an opportunity to study overseas. Multimedia plays an important role in current information society since multimedia efficiently represent perceptional data. Textual data, audio, images, and video are integrally called multimedia. Image is one of the most popular mediums. An image is a two dimensional signal stored as an array of pixels and symbolises visual information. " One image is worth thousand words" would be a sentence describing the information reflecting image meaning. It is clear that how to describe image meaning completely is difficult. This directly testifies that images effectively convey information rather than text. Moreover, many constituent complex sentences can be represented by an image. For example, novel computer operating systems such as windows 98 prefer to use graphical user interface such as icons instead of complicated commands. In addition, the use of images makes the deprivafion of descripfion problems to appear. For instance, advertising designers can visually demonstration their goods to prospective customers in order to query and generate orders. The problem of how to access and manage non-textual information is progressively apparent as expressed by the nature of image data, the storage requirements, the image representations, and the rapid growth of the use of image data. In this thesis, image INTRODUCTION meaning is obscure, and the access of image is problematic. Therefore, it is clear that the techniques achieve a high degree of compression, regardless of the information management perspective. Finally, users rapidly migrate from textual data services to multimedia services. This stems from two main reasons. Firstly, multimedia presentation perceptually conveys information. This makes an efficient communication process, and most information is comprehensive. This forms the new medium for information interchange. Secondly, a global network for information interchange is available. The Internet is the biggest network connecting the world. The Internet services such as world-wide-web encourage interchange of information among users. This is more so when users have easily understandable data stream like multimedia. Furthermore, the running cost of the Internet services is affordable for personal use. As indicated by the natural mean of multimedia, content-based retrieval of multimedia data is more appropriate. Supporting this heuristic, MPEG-7 standard aims to address this issue. From the information management perspective, retrieval of small data sets from a large collection of data streams is a central idea. Tremendous data streams are a significant impediment for retrieval. Fortunately, existing compression techniques obtain a high degree of compression. However, again, the techniques only support efficiency of storage and transmission management. They are developed based on representation models. The models include, (i) pixel intensities, (ii) pixels grouping to vectors, (iii) two dimensional grouping of pixels into regions occupying stationarity, and (iv) two dimensional grouping into region of pixels retaining high level semantic scheme such as object. Content-based choice of features. For instance, sub-string matching [67] is appropriate for keywords. The success of any feature set in measuring similarity depends more or less on the use of an appropriate metric. From the literature, existing image databases are specific to applications and are not supportive to content-based retrieval. Image retrieval in most of the systems is equivalent to content-based retrieval systems. Moreover, design criteria for retrieval systems are highlighted. Finally, a retrieval model is proposed for further experiments. Chapter 3 presents three image features that will be used for retrieval and discusses their characteristics. As similarity measures, the image metrics corresponding to the features are also briefly described. Recent studies indicate that content-based retrieval is more appropriate for multimedia data streams. From these studies, the solutions are not completely provided. Content-based retrieval relates to at least three aspects, namely the retrieval, database, and data representation aspects. This thesis aims to address the retrieval aspect. Conceptual issues of content-based retrieval are studied. An image model that is persuasive to content-based retrieval is proposed. From the retrieval point of view, the thesis nominates three features, namely the Zemike moments, the Fourier descriptors, and the wavelet descriptors. The features are extracted

Ambient Assisted Living (AAL) is trying to respond to the ageing process of the population, aiming at the extension of the time in which older people can live in their home environment and giving support to them and to their families. This paper highlights the importance of Artificial Intelligence to provide intelligent and adaptive solutions to assist an elderly patient with his daily activities, while his health conditions are monitored, in order to ensure his security. Thus a Multi-Agent architecture, based on Belief-Desire-Intention (BDI) paradigm, is proposed, modelling an AAL system. The adoption of BDI paradigm tries to represent the behaviours and the tasks of a human caregiver. The paper presents a software implementation of the proposed architecture: the Virtual Carer. By means of a simulation scenario, the main features and capabilities of the system are shown.

The ageing process of population of the First World countries is increasing the interest towards solutions to improve the quality of life of elderly or disabled people and their families, providing an economically sustainable healthcare. Thus Ambient Assisted Living (AAL) and Ambient Intelligence (AmI) are moving towards the Artificial Intelligence field to develop modular, adaptable and intelligent systems to cope with changing needs that characterize the life of people with chronic diseases. These systems should find smart and simple ways to improve the patient quality of life, enhancing at the same time the communication between the system, the assisted person, his family and the medical staff, acting also as a filter of information that provides useful and significant data. This paper introduces the basic architecture of an expert system for AAL: the Virtual Carer (VC). It has to be understood as an IT system modelling a distributed, reliable and modular sensor network composed by biometric and ambient sensors, being able to communicate with an assisted person, to monitor his health conditions and to control the environment around him. The main goal of the system is to help an elderly patient with his daily activities ensuring his security. The proposed system is based on a multi-agent architecture to ensure its flexibility and interoperability: new devices or sensors can be added simply adding new agents, thanks to the standardization of agent communication. The system includes also a reasoning part based on Belief-Desire-Intention (BDI) paradigm, trying to model the behaviours of a human caregiver.

In this paper a computational model is presented for how a desire triggers responses and feelings. The model shows how these feelings can be biased, for example due to addicting experiences in the past. Both the strength of a response and of the associated feeling result from a converging dynamic pattern modelled by reciprocal causal interactions between the two. The model has been used to conduct a number of simulation experiments under varying circumstances. Moreover, it has been evaluated by formal analysis of emerging patterns entailed by the model. Furthermore, it has been pointed out how the computational model can be applied within an ambient agent system supporting a human in not being tempted. In a simple example scenario it is shown such an ambient agent system is able to predict and assess a human's desire state, and use this assessment to suggest alternatives to avoid falling for certain temptations.

In this paper, we discuss the problem of UI adaptation to the context of use. To address this problem, we propose to mix declarative languages as promoted in Model Driven Engineering (MDE) with a “code-centric” approach where pieces of code are encapsulated as service-oriented components (SOA), all of this within a unified software framework that blurs the distinction

The advent and adoption of internet-based social networking has significantly altered our daily lives. The educational community has taken notice of the positive aspects of social networking such as creation of blogs and to support groups of system designers going through the same challenges and difficulties. This paper introduces a social networking framework for collaborative education, design and modeling of the next generation of smarter products and services. Human behavior modeling in social networking application aims to ensure that human considerations for learners and designers have a prominent place in the integrated design and development of sustainable, smarter products throughout the total system lifecycle. Social networks blend self-directed learning and prescribed, existing information. The selfdirected element creates interest within a learner and the ability to access existing information facilitates its transfer, and eventual retention of knowledge acquired. In conclusion, this research paper introduces the application of social-networking to design and modeling of products and services and provides a novel technology for facilitating the understanding of complex human behavior and to better identify crucial user needs.

Abstract. As technology develops, people are using an ever broader range of heterogeneous ICT (Information and Communication Technology) devices and network-based services. New areas of research, such as pervasive computing and Ambient Intelligence, will further increase the diversity of the devices and services with which users have to deal. The result is an enormous burden of complexity for users, service providers and network operators. This creates obstacles to effective exploitation and acceptance of ...

The context-aware systems allow the adaptation of the environment, which depends on the values perceived by sensors and on the users' preferences detected from user's behavior. These systems need to adapt at the users' preferences change so as to generate dynamic systems that improve their models while they are running. This work proposes a multi-agent system which allows to learn and to anticipate users' behavior and to adapt at the change into a dynamic mode, and where storing information from the previous cases during successive execution is no more compulsory.

The evolution of ambient intelligence systems has allowed for the development of adaptable systems. These systems trace user's habits in an automatic way and act accordingly, resulting in a context aware system. The goal is to make these systems adaptable to the user's environment, without the need for their direct interaction. This paper proposes a system that can learn from users' behavior. In order for the system to perform effectively, an adaptable multi agent system is proposed and the results are compared with the use of several classifiers.

This paper tackles the issue of ambient systems adaptation to users' needs while the environment and users' preferences evolve continuously. We propose the adaptive multiagent system Amadeus whose goal is to learn from users' actions and contexts how to perform actions on behalf of the users in similar contexts. However, considering the possible changes of users preferences, a previously learnt behaviour may become misfit. So, Amadeus must be able to observe if its actions on the system are contradicted by the users or not, without requiring any explicit feedback. The aim of this paper is to present the introspection capabilities of Amadeus in order to detect users contradictions and to self-adapt its behaviour at runtime. These mechanisms are then evaluated through a case study.

Roboethics is a recently developed field of applied ethics which deals with the ethical aspects of technologies such as robots, ambient intelligence, direct neural interfaces and invasive nano-devices and intelligent soft bots. In this article we look specifically at the issue of (moral) responsibility in artificial intelligent systems. We argue for a pragmatic approach, where responsibility is seen as a social regulatory mechanism. We claim that having a system which takes care of certain tasks intelligently, learning from experience and making autonomous decisions gives us reasons to talk about a system (an artifact) as being "responsible" for a task. No doubt, technology is morally significant for humans, so the "responsibility for a task" with moral consequences could be seen as moral responsibility. Intelligent systems can be seen as parts of socio-technological systems with distributed responsibilities, where responsible (moral) agency is a matter of degree. Knowing that all possible abnormal conditions of a system operation can never be predicted, and no system can ever be tested for all possible situations of its use, the responsibility of a producer is to assure proper functioning of a system under reasonably foreseeable circumstances. Additional safety measures must however be in place in order to mitigate the consequences of an accident. The socio-technological system aimed at assuring a beneficial deployment of intelligent systems has several functional responsibility feedback loops which must function properly: the awareness and procedures for handling of risks and responsibilities on the side of designers, producers, implementers and maintenance personnel as well as the understanding of society at large of the values and dangers of intelligent technology. The basic precondition for developing of this socio-technological control system is education of engineers in ethics and keeping alive the democratic debate on the preferences about future society.

The VICOM (Virtual Immersive COMmunications) project is a three-year project funded by the Italian Ministry of Instruction University and Research aiming at investigating innovative communication paradigms. The project represents a wide coordinated effort focused on integration of immersive and wireless technologies in view of the fourth generation of mobile communications. The main goal of the project consists of the design of a wideband system architecture for immersive services and of its validation through two distributed large test-beds. Starting from VICOM ongoing experiences some future challenges and objectives for the future situated and autonomic communications technologies are envisaged in the paper.

In this demonstration, we will show a context-aware information system intended for mobile users. The demonstration involves special-purpose hardware devices, called 'context tags', which can work with mobile devices such as mobile phones, to provide ambient information to users on the move. Key to the framework is special support for content service providers, through software that allows existing content to be delivered seamlessly to mobile devices, as and when it is needed by users. The demonstration will show how these components work together to provide an effective ambient information system for mobile users.

Recent advances of computing and networking technology have shifted computing convention from stationary to mobile. A forecast conducted by Gartner indicates that global mobile handset sales will increase exponentially in 2008 by 10% from 1.3 billion units in 2007 [13]. Furthermore, the popularity of wireless networking topology including both Wi-Fi access point and cellular mobile telecommunications enables users with constant access to online connection and further intensifies the demand of mobile devices. This provides the fundamental elements for creating ubiquitous environments of computing, networking, and interfacing that is both aware of and reactive to the presence of people. Such an environment is defined as Ambient Intelligence (AmI). Existing approaches that attempt to understand AmI environment mainly focus on how to seamlessly integrate hardware, i.e. mobile device and sensors, into human society and intelligently provide personalized knowledge and services. This has, however, left many essential issues unanswered, especially in regards to the integrity and performance of such dynamically distributed environments. In this chapter, we formulate a generic framework in which an AmI environment is generalized to consist only of users with devices, hosts where services are provided, and directory servers that act as information desks to users and hosts. In the proposed framework, the mobile agent notion is utilized to provide autonomous reasoning, learning, mobility, and collaboration features to construct AmI systems. Performance issues of load balancing and communication overhead in such a framework are then examined and analyzed against existing AmI techniques. Because of the dynamicity posed by AmI environments and the complexity of migrations and communications among agents, hosts and directory servers, it is necessary to provide a mechanism to warrant the accuracy and enhance the reliability of such an environment. Software verification employs formal methods of mathematically provable formulations to perform program analysis and model checking. Thus, general formulation using π-calculus is included to model the proposed approach to provide verification mechanism. In conclusion, the implementation of the proposed framework simulation using NetLogo is currently ongoing to visually demonstrate its feasibility. Some implementation details are included to stimulate comprehension of this framework.

The purpose of this paper is to introduce an innovative framework for implementation of ambient intelligence (AmI) environments. Compared to the existing state-of-the-art approaches, this framework creates a more decentralized and distributed AmI environment. In addition, the proposed approach is not limited to one specific domain, unlike many others. The openness of the presented architecture allows it to support a variety of devices ranged from small-embedded sensors to complex computing facilities. Finally, given that this approach is formulated based on multi-agent standard concepts, it can be easily implemented as add-on for existing software agent platforms to achieve rapid deployment. Implications for the development of this framework and future directions are discussed.

Several research programs are tackling the use of Wireless Sensor Networks (WSN) at specific fields, such as e-Health, e-Inclusion or e-Sport. This is the case of the project "Ambient Intelligence Systems Support for Athletes with Specific Profiles", which intends to assist athletes in their training. In this paper, the main developments and outcomes from this project are described. The architecture of the system comprises a WSN deployed in the training area which provides communication with athletes' mobile equipments, performs location tasks, and harvests environmental data (wind speed, temperature, etc.). Athletes are equipped with a monitoring unit which obtains data from their training (pulse, speed, etc.). Besides, a decision engine combines these real-time data together with static information about the training field, and from the athlete, to direct athletes' training to fulfill some specific goal. A prototype is presented in this work for a cross country running scenario, where the objective is to maintain the heart rate (HR) of the runner in a target range. For each track, the environmental conditions (temperature of the next track), the current athlete condition (HR), and the intrinsic difficulty of the track (slopes) influence the performance of the athlete. The decision engine, implemented by means of (m, s)-splines interpolation, estimates the future HR and selects the best track in each fork of the circuit. This method achieves a success ratio in the order of 80%. Indeed, results demonstrate that if environmental information is not take into account to derive training orders, the success ratio is reduced notably.

This work summarizes the state of research in universal design in ambient intelligence environments. We provide a detailed background, specify relevant research areas, and review research in these areas. We discuss the findings and put them into perspective with regard to universal design and accessibility in smart environments, and point out research shortcomings concerning ambient intelligence and hybrid interactions. Our findings show that the majority of related work needs stronger emphasis on aspects related to universal design in general; universal design in ambient intelligence; universal design in multimodal interactions; and universal design in security, privacy, and other ethical aspects of smart environments. Keywords—Universal design; ambient intelligence; multimodal interaction

In recent years, the widespread diffusion of pervasive sensing devices and the increasing need for reducing energy consumption have encouraged research in the energy-aware management of smart environments. Following this direction, this paper proposes a hybrid intelligent system which exploits a fog-based architecture to achieve energy efficiency in smart buildings. Our proposal combines reactive intelligence, for quick adaptation to the ever-changing environment, and deliberative intelligence, for performing complex learning and optimization. Such hybrid nature allows our system to be adaptive, by reacting in real time to relevant events occurring in the environment and, at the same time, to constantly improve its performance by learning users' needs. The effectiveness of our approach is validated in the application scenario of a smart home by extensive experiments on real sensor traces. Experimental results show that our system achieves substantial energy savings in the management of a smart environment, whilst satisfying users' needs and preferences.

Ageing society and individuals with disabilities faces numerous challenges in performing simple tasks in Activities of Daily Living (ADLs), [1]. ADLs represent the everyday tasks people usually need to be able to independently accomplish. Nowadays caring of elderly people becomes more and more important. Performance of ADLs in the long term view can be considered as a serious concern, especially when dealing with individuals who require extreme caregiver assistance. The objective of this paper is to introduce an Ambient Intelligence [2], real-scale home environment implementation, embedded with sensors and actuators, which enhances the independence and autonomy of the individuals upon performing ADLs. A real 1:1 scale experimental flat has been design and developed (Figure 1), in the authors experimental laboratory. A study was made to be able to determine the actual needs, required services, and functionality of the proposed augmented environment. In order to allow a high quality service delivery, a mobile autonomous rover [3] was introduced to act as the main humanmachine interface between the user and the distributed robotic systems and actuators. The mobile rover was wirelessly interfaced with the distributed intelligence to allow efficient interaction with the user, either passively, by vocal commands issued by the user, or adaptively, by autonomously navigating into the home environment. Moreover the Leap Motion hand gesture driven controller [4] is used as an intuitive user interface, which allows submillimeter accuracy capabilities, interfaced to a Jaco 6-Degrees of freedom robotic arm [5]. Many robotic systems have been designed and produced for assisting ADLs, in order to compensate this loss of mobility. However most of these solution are operated via a keyboard or joystick, which according to the complexity of the robotic system, require a series of configurations and mode selection routines in order to allow a specific trajectory path to be implemented. The authors developed new intuitive interfaces to operate such devices, in order to reduce the involved operation complexity. The user is able with simple gestures to operate and control complex robotic manipulators and mobile robots, without requiring the use of a joystick or a keyboard. The resulting accuracy and evaluated performance of the implemented interfaces, allow error free, continuous, and adaptive interaction between user queries and actuated environment responses.

Smart home is an adaptive system that controls house to make household members' life easier. There are plenty of not yet researched issues concerning this concept, such as the ability to update the architecture during system runtime. This paper presents a solution how to design such system on the JADE framework. To solve the problem of on-the-fly update, each agent needs to work in an OSGi environment. The proposal of new architecture is evaluated by means of the use case technique.

La thèse de doctorat, comme la plupart des activités de recherche, ne peut pas être seulement vue comme une tâche individuelle, l'aspect collaboration avec une équipe est fondamental. Lors de ces trois dernières années d'études qui ont représenté pour moi une des étapes les plus passionnantes de ma vie, j'ai eu l'opportunité d'être entouré de personnes très talentueuses dont les conseils m'ont beaucoup apporté, me permettant non seulement de conclure mes études de doctorat, mais aussi de donner un cap à ma carrière scientifique. En premier lieu, je tiens à remercier Michel Vacher et François Portet pour leur soutien et leur patience comme directeurs de thèse. Leur expérience et leur motivation ont rendu possible les contributions scientifiques qui constituent l'ossature de cette thèse. Je remercie également les autres membres de l'équipe GETALP : lorsque j'exposais l'état de mes recherches lors des réunions animés par Laurent Besacier, les retours qui m'étaient faits m'ont beaucoup aidé à avancer. Je voudrais aussi remercier Benjamin Lecouteux avec qui j'ai échangé de nombreuses idées. L'expérience d'Anthony Fleury m'a été précieuse et les discussions que nous avons eues ont toujours été fructueuses, je tiens à le remercier. J'aimerais remercier Arnaud Martin et François Charpillet pour avoir accepté de rapporter ma thèse. Je remercie egalement Patrick Reignier qui a accepté de présider le jury, ainsi que Christel Vrain, et Alain Anfosso pour leur participation au jury comme examinateurs. Je remercie Brigitte Meillon, Nicolas Bonnefond, et Mathieu Gallisot pour leur aide lors des expérimentations dans l'appartement Domus. Un grand merci à l'ensemble de doctorants de l'équipe Getalp : Fréderic, David, Sarah, Marwen, Quang et les doctorants de l'équipe MRIM : Andrew, Nadia, Mateusz, Johan, Path qui ont crée une bonne ambiance de travail et permis des moments de convivialité très agréables. Je remercie finalement Christelle pour son soutien et sa compréhension lors de la finalisation de cette thèse.

Home energy management is becoming increasingly important and, though there are a plethora of tools for accessing energy consumption data, few provide concrete insights that can directly help users manage demand. Mechanisms that enable a user to draw connections between activities and energy consumption by attaching contextual labels to energy events are a promising step; however solutions for collecting annotations from users can be error prone or intrusive. This work presents a system for collecting in situ annotations using a smartphone application coupled with an off-the-shelf home energy measurement infrastructure. We use a novel power profiling approach to identify important energy consumption events and solicit contextual annotations from the user via a push notification sent to a smartphone. Using a five-week study performed in five homes, we show that our power profiling approach can identify a significant percentage of important energy consumption events using a very small number of monitored devices. We were able to collect an average of over 2 annotations per day and while users provided a wide range of annotations, the motivation to provide annotations varied across subjects.

Home energy management is becoming increasingly important and, though there are a plethora of tools for accessing energy consumption data, few provide concrete insights that can directly help users manage demand. Mechanisms that enable a user to draw connections between activities and energy consumption by attaching contextual labels to energy events are a promising step; however solutions for collecting annotations from users can be error prone or intrusive. This work presents a system for collecting in situ annotations using a smartphone application coupled with an off-the-shelf home energy measurement infrastructure. We use a novel power profiling approach to identify important energy consumption events and solicit contextual annotations from the user via a push notification sent to a smartphone. Using a five-week study performed in five homes, we show that our power profiling approach can identify a significant percentage of important energy consumption events using a very small number of monitored devices. We were able to collect an average of over 2 annotations per day and while users provided a wide range of annotations, the motivation to provide annotations varied across subjects.

The deployment and use of Artificially Intelligent tools and techniques has proved to be effective and convenient, rendering them highly popular and desirable within any application. The benefits that such smart services provide are increasingly becoming popular within virtual learning environments as educators and learners continue to revert to such online portals during the academic process of higher education programmes. However, numerous educators have been sceptical of the inferences made by the machine learning techniques that implicitly at the background of the learning environment are monitoring the learner and collecting data to optimize the educational process as well as the learning experience. In an effort to enhance such a credibility we present an intelligent learning environment that justifies its decisions and propositions through an explainable interface that tracks back its conclusions to reasonable and plausible justifications. In this paper we present our researc...

This paper presents a distributed surveillance system to improve safety and security in public transport environments such as metropolitan railways, airports etc. The system is constructed by a set of modules that provide additional intelligence to detect potentially dangerous situations and provide more timely prevention and management of events. These modules include CCTV video processing, audio detection, mobile smart cards, a range of other possible sensors and a supervisory front-end. The system design uses ideas from control engineering and distributed network communication resulting in a CORBA based architecture and XML messaging. The system components have been implemented and integrated in two major successful trials in metropolitan railway stations (London and Paris) and at Newcastle International Airport, as part of a major EU-funded project PRISMATICA ("PRo-active Integrated systems for Security Management by Technological, Institutional and Communication Assistance").

The utilization of deep learning and transfer learning methodologies in the field of image classification has led to substantial advancements across a range of applications. This research paper presents an evaluation of the VGG16 convolutional neural network, enhanced through transfer learning and regularization techniques, for the specific task of sheep breed classification. The model was fine-tuned on a dataset of six sheep breeds, using preprocessing techniques like resizing, normalization, and augmentation. Architectural enhancements such as batch normalization, dropout, and dense layers helped reduce overfitting and improve generalization. A dropout rate of 0.5 with a batch size of 16 achieved the highest test accuracy of 80.00%. Higher dropout rates (e.g., 0.8) resulted in underfitting and lower performance. Overall, the use of transfer learning and dropout regularization significantly improved classification accuracy.

The evaluation of applications or systems within dynamic environments is complex. The existence of multiple hardware and software items which share the same space can provoke concurrency issues and result in erratic interactions. A sudden change within the environment can result is dramatic changes both to the user and application itself which can pass unnoticed in traditional evaluation methodologies. To verify if a component is compatible with a given environment is of paramount importance for areas like pervasive computing, ambient intelligence or ambient assisted living (AAL). In this paper, a semi-automatic platform for evaluation is presented and integrated with a TeleRehabilitation system in an AAL scenario to enhance evaluation. Preliminary results show the advantages of the platform in comparison with typical observation solutions mainly in terms of achieved data and overall ease of use.