Digital twins are becoming increasingly relevant in various industries, particularly in the manufacturing sector. While general architectures exist, there is a need for specific solutions tailored to sectors with limited research, such as ceramics. This article presents a flexible digital twin architecture for quality control in the ceramics industry, specifically for tableware, combining Industry 5.0 requirements while remaining compatible with Industry 4.0. The architecture is composed of seven distinct entities and is designed for practical implementation in real-world settings. The solution’s flexibility and adaptability promote the digital transformation of ceramic manufacturing. In a case study, the system achieved a low-latency, real-time quality control system with 98.00% accuracy, a 97.29% F1 score for classification, and an average precision ([email protected]) of 73.10%. Furthermore, the physical workload of operators was reduced, enhancing human well-being, a fundamental component of Industry 5.0.

ABSTRACT This article presents a novel approach for real‐time posture recognition in monitoring scenarios, utilising a virtual camera simulated on a UAV within virtual environments. Leveraging the MediaPipe Pose library, key points of the body skeleton are extracted, focusing on a subset of 8 key points for computational efficiency. Through the integration of heuristic algorithms based on physical proportions of the human body, the proposed methodology provides accurate estimations of three distinct postures: lying, standing, and sitting. This heuristic‐based approach offers a computationally efficient alternative to traditional machine learning and deep learning methods, ensuring real‐time performance and scalability. The efficiency of the framework is demonstrated through experiments that show its potential applications in various fields, including healthcare, virtual reality, and human‐computer interaction. This approach achieved an average precision of 98.08% for virtual images. Success rates were 100%, 95.8%, and 98.9% for standing, sitting, and lying postures, respectively. Furthermore, the original classification model, which was tuned for virtual images, was tested on real images without any alteration to the parameter values. Its good performance demonstrates its potential for generalisation and application in diverse environments. Overall, this work contributes to the advancement of posture recognition technology, offering a versatile and accessible solution for posture analysis in dynamic monitoring environments.

Emotion recognition has gained popularity in recent years. However, it is still a very challenging task due to the complexity, variety and multi-modality of human emotion expression. Focusing on visual data, body language plays an important role in emotion recognition, although this topic remains under-explored. This paper presents a thorough study of emotion recognition using seven machine learning techniques applied to three different configurations of human skeletal landmarks. A first configuration uses a reduced set of body pose landmarks obtained by MediaPipe. A second configuration also uses MediaPipe, but focuses on obtaining hand pose landmarks. Finally, the third configuration selects a number of landmarks from the previous two configurations. Five basic emotions (joy, pride, anger, sadness and fear) from the multimodal GEMEP dataset are considered for comparison. Remarkably, the best accuracy for all configurations was greater than 99.5%. The highest accuracy was achieved by Random Forest (RF) classifier for the first two configurations (99.82% and 96.43% respectively), closely followed by Gradient Boosting (GB). The third configuration showed an accuracy of 98.81% by GB, followed by 97.62% by RF. It should be noted that our proposal achieves an improvement in accuracy of almost 4% compared to other approaches using the GEMEP dataset.

Research in various fields has highlighted the benefits of using extended reality (XR) in many applications. However, the use of musical instruments in this context remains largely unexplored. This article presents a systematic review in which the existing literature is compiled and analyzed to identify the current state of research. A systematic search of four electronic databases of articles published in English since 2010 was conducted. After merging duplicates, a total of 2849 unique records were identified, of which 74 were included in the analysis. The review reveals a diverse landscape of XR applications, ranging from music learning to rehabilitation therapies. Diverse interaction modalities and factors affecting the musician’s experience have been identified. XR technologies offer unprecedented opportunities for musicians; however, technological limitations pose challenges to creating fully satisfying experiences. This review is intended to guide future research that supports the development of new forms of musical expression in XR.

Social robots are increasingly being used in therapeutic contexts, especially as a complement in the therapy of children with Autism Spectrum Disorder (ASD). Because of this, the aim of this study is to understand how children with ASD perceive and interpret the gestures made by the robot Pepper versus human instructor, which can also be influenced by verbal communication. This study analyzes the impact of both conditions (verbal and nonverbal communication) and types of gestures (conversational and emotional) on gesture recognition through the study of the accuracy rate and examines the physiological responses of children with the Empatica E4 device. The results reveal that verbal communication is more accessible to children with ASD and neurotypicals (NT), with emotional gestures being more interpretable than conversational gestures. The Pepper robot was found to generate lower responses of emotional arousal compared to the human instructor in both ASD and neurotypical children. This study highlights the potential of robots like Pepper to support the communication skills of children with ASD, especially in structured and predictable nonverbal gestures. However, the findings also point to challenges, such as the need for more reliable robotic communication methods, and highlight the importance of changing interventions tailored to individual needs.

Areas, such as the identification of human activity, have accelerated thanks to the immense development of artificial intelligence (AI). However, the lack of data is a major obstacle to even faster progress. This is particularly true in computer vision, where training a model typically requires at least tens of thousands of images. Moreover, when the activity a researcher is interested in is far from the usual, such as falls, it is difficult to have a sufficiently large dataset. An example of this could be the identification of people suffering from a heart attack. In this sense, this work proposes a novel approach that relies on generative models to extend image datasets, adapting them to generate more domain-relevant images. To this end, a refinement to stable diffusion models was performed using low-rank adaptation. A dataset of 100 images of individuals simulating infarct situations and neutral poses was created, annotated, and used. The images generated with the adapted models were evaluated using learned perceptual image patch similarity to test their closeness to the target scenario. The results obtained demonstrate the potential of synthetic datasets, and in particular the strategy proposed here, to overcome data sparsity in AI-based applications. This approach can not only be more cost-effective than building a dataset in the traditional way, but also reduces the ethical concerns of its applicability in smart environments, health monitoring, and anomaly detection. In fact, all data are owned by the researcher and can be added and modified at any time without requiring additional permissions, streamlining their research.

The recent proliferation of the Industrial Internet of Things has revealed the potential of Low-Power Wide-Area Networks as a complementary solution to cellular technologies. In this context, the LoRaWAN standard has already been consolidated as one of the most extended technologies in academia and industry for lightweight machine-type communications under negligible energy and cost. As LoRaWAN’s Aloha-like nature is known to hinder its reliability, especially under high-traffic and large-scale deployments, numerous time-slotted approaches have been presented as a means to schedule LoRa transmissions accordingly. However, the online allocation of resources based on application constraints has received scant attention in the literature, despite having proved to be significant in real-world deployments. To shed light on this question, this paper proposes a multi-agent approach to efficient resource allocation in multi-SF LoRaWAN networks, addressing architecture design, logic implementation and scalability-oriented evaluation. The integration of agents in the system resulted in network-size improvements of up to 21.6% and 66.7% (for nearby or scatter node distributions within the gateway, respectively). The work provides a set of learned lessons regarding slot-length computation and end-node allocation strategies enabling large-scale collision-free channel access in LoRaWAN networks.

Facial affect recognition is a critical skill in human interactions that is often impaired in psychiatric disorders. To address this challenge, tests have been developed to measure and train this skill. Recently, virtual human (VH) and virtual reality (VR) technologies have emerged as novel tools for this purpose. This study investigates the unique contributions of different factors in the communication and perception of emotions conveyed by VHs. Specifically, it examines the effects of the use of action units (AUs) in virtual faces, the positioning of the VH (frontal or mid-profile), and the level of immersion in the VR environment (desktop screen versus immersive VR). Thirty-six healthy subjects participated in each condition. Dynamic virtual faces (DVFs), VHs with facial animations, were used to represent the six basic emotions and the neutral expression. The results highlight the important role of the accurate implementation of AUs in virtual faces for emotion recognition. Furthermore, it is observed that frontal views outperform mid-profile views in both test conditions, while immersive VR shows a slight improvement in emotion recognition. This study provides novel insights into the influence of these factors on emotion perception and advances the understanding and application of these technologies for effective facial emotion recognition training.

The identification of the emotional states corresponding to the four quadrants of the valence/arousal space has been widely analyzed in the scientific literature by means of multiple techniques. Nevertheless, most of these methods were based on the assessment of each brain region separately, without considering the possible interactions among different areas. In order to study these interconnections, this study computes for the first time the functional connectivity metric called cross-sample entropy for the analysis of the brain synchronization in four groups of emotions from electroencephalographic signals. Outcomes reported a strong synchronization in the interconnections among central, parietal and occipital areas, while the interactions between left frontal and temporal structures with the rest of brain regions presented the lowest coordination. These differences were statistically significant for the four groups of emotions. All emotions were simultaneously classified with a 95.43% of accuracy, overcoming the results reported in previous studies. Moreover, the differences between high and low levels of valence and arousal, taking into account the state of the counterpart dimension, also provided notable findings about the degree of synchronization in the brain within different emotional conditions and the possible implications of these outcomes from a psychophysiological point of view.

The assessment of physiological signals such as the electroencephalography (EEG) has become a key point in the research area of emotion detection. This study compares the performance of two EEG devices, a low-cost brain–computer interface (BCI) (Emotiv EPOC+) and a high-end EEG (BrainVision), for the detection of four emotional conditions over 20 participants. For that purpose, signals were acquired with both devices under the same experimental procedure, and a comparison was made under three different scenarios, according to the number of channels selected and the sampling frequency of the signals analyzed. A total of 16 statistical, spectral and entropy features were extracted from the EEG recordings. A statistical analysis revealed a major number of statistically significant features for the high-end EEG than the BCI device under the three comparative scenarios. In addition, different machine learning algorithms were used for evaluating the classification performance of the features extracted from high-end EEG and low-cost BCI in each scenario. Artificial neural networks reported the best performance for both devices with an F[Formula: see text]-score of 75.08% for BCI and 98.78% for EEG. Although the professional EEG outcomes were higher than the low-cost BCI ones, both devices demonstrated a notable performance for the classification of the four emotional conditions.

The recognition of facial expression of emotions in others is essential in daily social interactions. The different areas of the face play different roles in decoding each emotion. To find out which ones are more important, the traditional approach has been to use eye-tracking devices with static pictures to capture which parts of the face people are looking at when decoding emotions. However, the ecological validity of this approach is limited because, unlike in real life, there is no movement in the face that can be used to identify the emotion. The use of virtual reality technology opens the door to new experiences in which the users perceive that they are in front of dynamic virtual humans. Therefore, our main aim is to investigate whether the user's immersion in a virtual environment influences the way dynamic virtual faces are visually scanned when decoding emotions. An experiment involving 74 healthy participants was carried out. The results obtained are consistent with previous works. Having confirmed the correct functioning of our solution, it is our intention to study whether emotion recognition deficits in patients with neuropsychiatric disorders are related to the way they visually scan faces.

While the Internet of Things continues to grow, the LoRaWAN standard is generating special interest due to its open-source nature, ultralow-power consumption and long-range connectivity. Recent works have explored the challenges of implementing LoRaWAN, with scalability being considered one of the major bottlenecks imposed by its Aloha-based medium access control (MAC) layer. Despite much on-going research on LoRaWAN scheduling aimed at alleviating this concern, experimental approaches are rarely found in the literature. In this work, we describe the steps taken and the technical issues overcome to move from a low-overhead synchronization and scheduling concept to its real-world implementation on top of LoRaWAN Class A. Accordingly, an end-to-end architecture was designed and deployed on top of STM32L0 MCUs, which communicate with a central entity responsible for providing synchronization metrics and allocating transmission slots on demand. The clock drift of devices was measured in a temperature-controlled chamber, which served as a basis to define slot lengths in the network. As a result, an operational end-to-end system was implemented and evaluated for different setup scenarios, with 10-ms accuracy being achieved. Our experimental results show significant improvements in packet delivery ratios with respect to Aloha-based setups, especially under high network loads (up to 29% for SF12), thereby demonstrating the feasibility of the presented approach.

Negative stress, or distress, represents a serious problem in advanced societies given its adverse consequences for health. Many studies have focused on the detection of distress from physiological signals such as the electroencephalogram (EEG). To this respect, the combination of regularity-based quadratic sample entropy (QSampEn) and symbolic amplitude-aware permutation entropy (AAPE) has reported valuable outcomes in distress recognition. In the present work, the recently introduced symbolic metric called dispersion entropy (DispEn) is applied for the first time to the same problem. Statistically significant results reported by the single metric have demonstrated its capability for calm and distress detection. Furthermore, relevant differences have been found between the combination of QSampEn with either AAPE or DispEn, finding that the assessment of ordinal and dispersion patterns leads to distinct and complementary outcomes. Finally, the combination of the three entropy metrics has considerably overcome the results ever reported by other indices in similar studies.

Electroencephalographic (EEG) recordings are receiving growing attention in the field of emotion recognition, since they monitor the brain's first response to an external stimulus. Traditionally, EEG signals have been studied from a linear viewpoint by means of statistical and frequency features. Nevertheless, given that the brain follows a completely nonlinear and nonstationary behavior, linear metrics present certain important limitations. In this sense, the use of nonlinear methods has recently revealed new information that may help to understand how the brain works under a series of emotional states. Hence, this paper summarizes the most recent works that have applied nonlinear methods in EEG signal analysis for emotion recognition. This paper also identifies some nonlinear indices that have not been employed yet in this research area.

Abstract This paper introduces a wearable hardware/software system specifically tailored to detect seven emotions (neutral, tenderness, amusement, anger, disgust, fear, and sadness) aimed at promoting health and wellness in older adults living alone at home. The complete software and hardware architectures acquiring and processing electrodermal activity and photoplethysmography signals are introduced. The wearable emotion detection system is trained by eliciting the desired emotions on 39 older adults through a film mood induction procedure. Seventeen features are calculated on skin conductance response and heart rate variability data, grouped into five statistical, four temporal, and eight morphological features. Then, these features are used to run emotion classification considering support vector machines, decision trees, and quadratic discriminant analysis. In line with psychological findings, the results offer a global accuracy of 82% in negative emotion (anger, disgust, fear, and sadness) classification. For positive emotions (tenderness and amusement), also in conformity with previous psychological outcomes, amusement shows the highest ratio of hits (92%) but tenderness the lowest one (66%). These results demonstrate that our wearable emotion detection system can be used by ageing adults, especially for detecting negative emotions that usually damage health and wellness and lead to social isolation.

Early detection of stress condition is beneficial to prevent long-term mental illness like depression and anxiety. This paper introduces an accurate identification of stress/calm condition from electrodermal activity (EDA) signals. The acquisition of EDA signals from a commercial wearable as well as their storage and processing are presented. Several time-domain, frequency-domain and morphological features are extracted over the skin conductance response of the EDA signals. Afterwards, a classification is undergone by using several classical support vector machines (SVMs) and deep support vector machines (D-SVMs). In addition, several binary classifiers are also compared with SVMs in the stress/calm identification task. Moreover, a series of video clips evoking calm and stress conditions have been viewed by 147 volunteers in order to validate the classification results. The highest F1-score obtained for SVMs and D-SVMs are 83% and 92%, respectively. These results demonstrate that not only classical SVMs are appropriate for classification of biomarker signals, but D-SVMs are very competitive in comparison to other classification techniques. In addition, the results have enabled drawing useful considerations for the future use of SVMs and D-SVMs in the specific case of stress/calm identification.

Artificial intelligence and all its supporting tools, e.g. machine and deep learning in computational intelligence-based systems, are rebuilding our society (economy, education, life-style, etc.) and promising a new era for the social welfare state. In this paper we summarize recent advances in data science and artificial intelligence within the interplay between natural and artificial computation. A review of recent works published in the latter field and the state the art are summarized in a comprehensive and self-contained way to provide a baseline framework for the international community in artificial intelligence. Moreover, this paper aims to provide a complete analysis and some relevant discussions of the current trends and insights within several theoretical and application fields covered in the essay, from theoretical models in artificial intelligence and machine learning to the most prospective applications in robotics, neuroscience, brain computer interfaces, medicine and society, in general.

Automatic identification of negative stress is an unresolved challenge that has received great attention in the last few years. Many studies have analyzed electroencephalographic (EEG) recordings to gain new insights about how the brain reacts to both short- and long-term stressful stimuli. Although most of them have only considered linear methods, the heterogeneity and complexity of the brain has recently motivated an increasing use of nonlinear metrics. Nonetheless, brain dynamics reflected in EEG recordings often exhibit a multiscale nature and no study dealing with this aspect has been developed yet. Hence, in this work two nonlinear indices for quantifying regularity and predictability of time series from several time scales are studied for the first time to discern between visually elicited emotional states of calmness and negative stress. The obtained results have revealed the maximum discriminant ability of 86.35% for the second time scale, thus suggesting that brain dynamics triggered by negative stress can be more clearly assessed after removal of some fast temporal oscillations. Moreover, both metrics have also been able to report complementary information for some brain areas.

For the sake of establishing the neural correlates of phrase quadrature perception in harmonic rhythm, a musical experiment has been designed to induce music-evoked stimuli related to one important aspect of harmonic rhythm, namely the phrase quadrature. Brain activity is translated to action through electroencephalography (EEG) by using a brain-computer interface. The power spectral value of each EEG channel is estimated to obtain how power variance distributes as a function of frequency. The results of processing the acquired signals are in line with previous studies that use different musical parameters to induce emotions. Indeed, our experiment shows statistical differences in theta and alpha bands between the fulfillment and break of phrase quadrature, an important cue of harmonic rhythm, in two classical sonatas.

This paper seeks to determine how the overlap of several infrared beams affects the tracked position of the user, depending on the angle of incidence of light, distance to the target, distance between sensors, and the number of capture devices used. We also try to show that under ideal conditions using several Kinect sensors increases the precision of the data collected. The results obtained can be used in the design of telerehabilitation environments in which several RGB-D cameras are needed to improve precision or increase the tracking range. A numerical analysis of the results is included and comparisons are made with the results of other studies. Finally, we describe a system that implements intelligent methods for the rehabilitation of patients based on the results of the tests carried out.

Human-agent societies refer to applications where virtual agents and humans coexist and interact transparently into a fully integrated environment. One of the most important aspects in this kind of applications is including emotional states of the agents (humans or not) in the decision-making process. In this sense, this paper presents the applicability of the JaCalIVE (Jason Cartago implemented intelligent virtual environment) framework for developing this kind of society. Specifically, the paper presents an ambient intelligence application where humans are immersed into a system that extracts and analyzes the emotional state of a human group. A social emotional model is employed to try to maximize the welfare of those humans by playing the most appropriate music in every moment.

Abstract This paper addresses the challenging problem of disorientation of elderly people living at home. In order to detect confusion, we monitor the behaviour of the elderly and identify actions that appear alarming in a sensorized and video‐controlled smart environment. In the past, our research has focused on identifying situations, activities and interactions between various actors based on user‐understandable models. This work addresses the development of a simulation tool capable of synthesizing sensor data and low‐level/medium‐level scene events. The tool is of great interest with regard to the design and configuration of an elderly disorientation recognition system because it reduces the laborious and expensive need for experimentation with real devices. We integrate this proposal in a comprehensive framework that distinguishes between a recognition line and a simulation line in a potentially continuous and closed cycle. The recognition line goes from a multisensory monitored scene to its semantic interpretation, which could be completed even with only the narration of the facts. In the opposite direction, the simulation line goes from the narration of a scene to its synthesis with the same semantic content into a 3D simulation and the corresponding sensor signals and low/medium events at specific location points.

Abstract This paper proposes the use of RGB‐D sensors in motor rehabilitation of patients suffering acquired brain injury. Indeed, RGB‐D sensors are promising in their use as assistive technologies in ambient assisted living and rehabilitation places. The paper describes the system developed for creating and monitoring the ambient assisted environment for rehabilitation. The proposal consists of four major tasks: workspace configuration, information integration, user identification, and user tracking. Another main concern in this paper is to assess the user experience of the environment and scenario. A qualitative evaluation of the proposal has been performed to find out how the stakeholders of the proposed RGB‐D assistive technologies environment (injured patients and therapists) use and perceive the ambient assisted rehabilitation room.

Human beings use spatial relations to describe many daily tasks in their language. For a mobile robot to be useful in daily life, it is necessary to have navigation algorithms capable of identifying and establishing spatial relations. To date in robotics, the navigation problem has been thoroughly researched as a task of guiding a robot from one spatial coordinate to another. Therefore, there is a difference in degree of abstraction between the language of human beings and the algorithms used in robot navigation. This article introduces a piece of research performed on the use of topological relations for the formalisation of spatial relations and navigation. So far, topological relations have been applied widely in geographical information systems and also in spatial logics. There are some proposals in robot navigation which use them for planning but there is no research about making decision in robot navigation. Our research focuses on decision-making methods to establish spatial relations. The main result is a new heuristic, called the Heuristic of Topological Qualitative Semantics (HTQS), which allows the identification and establishment of spatial relations decision-making from a set of actions. To demonstrate its effectiveness, HTQS has been implemented in the form of agents that can move in a two-dimensional virtual environment. HTQS opens a new door to designing algorithms for navigation based on the identification and establishment of spatial relations.

Abstract VigilAgent is a methodology for the development of agent‐oriented monitoring applications that uses agents as the key abstraction elements of the involved models. It has not been developed from scratch, but it reuses fragments from Prometheus and INGENIAS methodologies for modelling tasks and the ICARO framework for implementation purposes. As VigilAgent intends to automate as much as possible the development process, it exploits. Model transformation techniques are one of the key aspects of the model‐driven development approach. A model‐to‐model transformation is used to facilitate the interoperability between Prometheus and INGENIAS methodologies. Also, a model‐to‐text transformation is performed to generate ICARO code from the INGENIAS model. A case study based on access control is used to illustrate the fundamentals of the model‐to‐model and model‐to‐text transformations implemented in VigilAgent.

Intelligent monitoring for people assistance and safetyThis expert systems special issue on 'Intelligent Monitoring for People Assistance and Safety' contains the revised and extended best papers dealing with different issues concerning people assistance and safety through intelligent monitoring and activity interpretation, presented at 'IWINAC 2011: the fourth International Work-Conference on the Interplay between Natural and Artificial Computation'.People assistance and safety is a hot topic and of crucial importance in indoor environments such as homes, offices, hospitals and schools as well as in outdoor areas.Environments are increasingly well equipped with multiple sensing technologies that can monitor simple and complex activities and behaviours (Gascueña and Fernández-Caballero, 2011).Intelligent monitoring implies not only the analysis of the data captured from the various sensors but also their interpretation from the detection of the presence of certain events or actions previously defined (Rivas, Martínez-Tomás and Fernández-Caballero, 2011).From a historical perspective, it is acknowledged that the evolution of monitoring systems has gone through three generations.In the first generation , closedcircuit television analogue systems were used, which consisted of several cameras connected to a series of monitors.These systems do not process information and require a human operator to be permanently concentrated on analysing the situations observed on the monitors.However, in the second generation (1990)(1991)(1992)(1993)(1994)(1995)(1996)(1997)(1998)(1999)(2000), advances attained in digital video communication (e.g.digital compression, bandwidth reduction and robust transmission) were used to increase the efficiency of monitoring systems: closed-circuit television systems were combined with computer vision technology to process images automatically, in order to be proactive in the detection of alarm events during recording.These semiautomatic systems required a robust tracking and detection algorithm for behaviour analysis.Whereas these systems represented a clear improvement with respect to first generation systems by reducing the dependency on human operators to detect anomalous situations, their algorithms and techniques were responsible for triggering a high number of false positives.In the third generation (2000-today), a series of heterogeneous sensors (e.g.fixed cameras, pan-tilt-zoom (PTZ) cameras, audio sensors and RFID tags (radio-frequency identification) will be geographically distributed throughout the scenario to be observed.From the image processing point of view, these systems are based on distributed processing capabilities and the use of embedded signal processing devices to gain distributed scalability and robustness.The main problems that need to be solved in third generation

Complex emergent systems are known to be ill-managed because of their complex nature. This article introduces a novel interdisciplinary approach towards their study. In this sense, the DeciMaS methodological approach to mining and simulating data in complex information systems is introduced. The De ciMaS framework consists of three principal phases, preliminary domain and system analysis, system design and coding, and simulation and decision making. The framework offers a sequence of steps in order to support a domain expert who is not a specialist in data mining during the knowledge discovery process. With this aim a generalized structure of a decision support system (DSS) has been worked out. The DSS is virtually and logically organized into a three-leveled architecture. The first layer is dedicated to data retrieval, fusion and pre-processing, the second one discovers knowledge from data, and the third layer deals with making decisions and generating output information. Data mining is aimed to solve the following problems: association, classification, function approximation, and clustering. DeciMaS populates the second logical level of the DSS with agents which are aimed to complete these tasks. The agents use a wide range of data mining procedures that include approaches for estimation and prediction: regression analysis, artificial networks (ANNs), self-organizational methods, in particular, Group Method of Data Handling, and hybrid methods. The association task is solved with artificial neural networks. The ANNs are trained with different training algorithms such as backpropagation, resilient propagation and genetic algorithms. In order to assess the proposal an exhaustive experiment, designed to evaluate the possible harm caused by environmental contamination upon public health, is introduced in detail.

The exponential increase of home-bound persons thatlive alone and are in need of continuous monitoring requires newsolutions to current problems. Most of these cases presentillnesses, such as motor or psychological disabilities, that deprivethem of a normal living. Abnormal situations such asforgetfulness or falls are quite common and should be preventedor dealt with. This paper presents a system able to detectdangerous situations at home, such as falls, independently fromexisting environment conditions. The aim of the proposed systemis to proactively offer support to the citizen or to warn theemergency services when needed.

This paper introduces the technical description of ICT R&D project Fall Detection -- HOLDS. The purpose of this project is to quickly assist elderly people or people who have a cognitive or motor disability when they suffer a fall at outdoor and indoor spaces. The project is aimed at enhancing the safety of these target groups and allows the elderly and impaired to carry out a more normal lifestyle. The HOLDS project is based on currently rising technologies, namely wireless sensors & actuators networks (WSAN) combined with advanced image processing. The HOLDS system comprises two subsystems, accelerometer-based fall detection and computer-vision-based (visible and infrared) fall detection, which are collected in a central system.

The INT3-Horus framework, dedicated to intelligent monitoring and activity interpretation with special application in advanced surveillance systems, is introduced. INT3-Horus is presented in two parts. This paper introduces the first part which is dedicated to provide a general description of the approach. The following aspects of the proposal are highlighted: the framework is multisensory by nature and includes information fusion abilities; it is based on the model–view–controller paradigm; it is defined as a hybrid distributed system; it incorporates a Common Model that houses the data structures to support the exchange of information between levels of the framework. The currently available INT3-Horus processing levels are also described in this paper. The second part introduces the ontological model of the framework [14].

Vision-based fall detection is a challenging problem in pattern recognition. This paper introduces an approach to detect a fall as well as its type in color video sequences. Accumulative computation is used to segment the color videos for the sake of robustly detecting humans. The regions of interest of the segmented humans are examined through calculating geometrical and kinematic features. The fall indicators used as well as their fuzzy model are explained in detail. The fuzzy model has been tested for a wide number of static and dynamic falls.

The INT3-Horus framework, dedicated to intelligent monitoring and activity interpretation with special application in advanced surveillance systems, is introduced. INT3-Horus is presented in two parts. This paper introduces the second part of the description of the framework. Here we introduce the framework ontological model. The ontology is composed of a couple of classes, namely the Level Class and the DataType Class. The paper also describes the relations between both classes, as well as it introduces the notion of set of rules which determine the system functionality for a given domain. The first part introduces the general description of the framework [1].

Modern intelligent monitoring and interpretation systems manage several kinds of heterogeneous sensor networks and use outstanding segmentation and tracking algorithms. Monitoring has evolved from initial systems based on low resolution cameras, directly connected to a monitor, up to distributed systems where several sensors cooperate not only to track objects of interest but also to detect suspicious behaviors based on artificial intelligence techniques. In our opinion, frameworks are essential to provide design and implementation patterns for generating a widespread variety of monitoring and interpretation applications, allowing the interaction of different modules and the reuse of code. In this sense, this paper proposes the implementation of a multi sensory monitoring and interpretation framework based on the model-view-controller paradigm but extended to distributed intelligent systems.

This paper is related to ambient intelligence systems capable of locating and tracking humans. These are the first steps of a human-centered ambient intelligent system, ranging from data acquisition to robust tracking, for the purpose of interpreting human behaviors in monitored environments. The first objective is to improve human detection through the fusion of thermal-infrared and color video segmentation. On the level following to segmentation, the traditional tracking problems (e.g. occlusions, crossings, etc.) are faced. Finally, the use of several classifiers such as support-vector machines and artificial neural networks are proposed to enhance the tracking level. The work proposes a combination of both color and thermal infrared spectra in human tracking.

This work presents a multi-agent system for knowledge-based high-level event composition, which interprets activities, behaviour and situations semantically in a scenario with multi-sensory monitoring. A perception agent (plurisensory agent and visual agent)-based structure is presented. The agents process the sensor information and identify (agent decision system) significant changes in the monitored signals, which they send as simple events to the composition agent that searches for and identifies pre-defined patterns as higher-level semantic composed events. The structure has a methodology and a set of tools that facilitate its development and application to different fields without having to start from scratch. This creates an environment to develop knowledge-based systems generally for event composition. The application task of our work is surveillance, and event composition/inference examples are shown which characterize an alarming situation in the scene and resolve identification and tracking problems of people in the scenario being monitored.

Abstract: In this work the concept of computational agent is located within the methodological framework of levels and domains of description of a calculus in the context of different usual paradigms in Artificial Intelligence (symbolic, situated, connectionist, and hybrid). Emphasis in the computable aspects of agent theory is put, leaving open the possibility to the incorporation of other aspects that are still pure cognitive nomenclature without any computational counterpart of equivalent semantic richness. The ideas presented are shown as currently being implemented on semi‐automatic surveillance systems. A case study of a mobile robot application for the detection and following of humans is described.

As far as a surveillance system is always integrated in an environment it has to adapt to possible changes that can occur in it. For this reason, it is not sufficient to install a series of sensors along the facilities to be guarded, as any modifications enormously increment the amount of data to be interpreted. Also, eventual failures or sabotages to the sensors produce the collapse of the system, which is not convenient at all in a potentially dangerous environment. Thus, the natural evolution of these systems is the integration in a compact system of intelligent platforms, which are able or not of moving in the environment, which possess several and complementary sensor types, and which interpret the information of each sensor coherently to offer the platform itself a fair service of surveillance. Quality of service is notably increased when there are a sufficient number of platforms forming a compact multi-agent system (MAS). Moreover, this MAS can itself be a compact subsystem of a superior hierarchy MAS composed of several subsystems. This is what we denominate recursive or holonic multi-agent systems.

This paper introduces domain ontology to describe learning materials that compose a course, capable of providing adaptive e-learning environments and reusable educational resources. Two characteristics have been considered to describe each resource: (1) the most appropriate learning style and, (2) the most satisfactory hardware and software features of the used device

All books entitled "Learn ... with 1000 exercises" have in common the same basic principle. They aim to supply enough material to students so that they may better understand the studied subject, starting from their own practice. If there is no instructor who helps students during the reading of the book, the students will not be able to understand the subject, as the excessive amount of information provided in this kind of books does not enable learners to pursue the learning goals. There is a great boom in e-learning through the socalled Intelligent Tutoring Systems, excellent virtual instructors which guide their learners through the reading of such kinds of books and help their learners to classify all the exercises and recommend them which ones to solve first. Nowadays instructors and teachers are entrusted to produce these books and to classify all exercises, whatever implies an overload to teachers. In this work we introduce a scalable system that only requires teachers to write the questions and their answers. The system will classify and manage all the questions. So the teacher will obtain, with the minimal effort, hundreds of exercises at the end of the course (and for future courses) which will reinforce individually his students.

Along this document, we expose how user interface adaptivity principles may contribute to improve present Web learning systems. If we are able to establish some student features, based on the documents he has consulted, we can easily guide him through our Web site. In order to collect that information, the Web site considers the metadata associated with each document. Once this information has been obtained, the interface recommends the site documents that are the most tailored to the student's interests. Our proposal consists in mixing the main ideas inherent to user profiling, semantic Web, metadata, and ontologies. The framework supporting the initiative is also introduced in this paper.

In this paper, we introduce a project that validates the educational capabilities of a game. This game, called Prismaker, incorporates two versions: a physical version and a virtual version. Thus, we want to find out the real potentialities of games in learning processes and to evaluate a single game from two points of view: the physical version and the virtual version that is being developed.