Field Of Ambient Intelligence Research Articles

Using Graphs to Perform Effective Sensor-Based Human Activity Recognition in Smart Homes.

There has been a resurgence of applications focused on human activity recognition (HAR) in smart homes, especially in the field of ambient intelligence and assisted-living technologies. However, such applications present numerous significant challenges to any automated analysis system operating in the real world, such as variability, sparsity, and noise in sensor measurements. Although state-of-the-art HAR systems have made considerable strides in addressing some of these challenges, they suffer from a practical limitation: they require successful pre-segmentation of continuous sensor data streams prior to automated recognition, i.e., they assume that an oracle is present during deployment, and that it is capable of identifying time windows of interest across discrete sensor events. To overcome this limitation, we propose a novel graph-guided neural network approach that performs activity recognition by learning explicit co-firing relationships between sensors. We accomplish this by learning a more expressive graph structure representing the sensor network in a smart home in a data-driven manner. Our approach maps discrete input sensor measurements to a feature space through the application of attention mechanisms and hierarchical pooling of node embeddings. We demonstrate the effectiveness of our proposed approach by conducting several experiments on CASAS datasets, showing that the resulting graph-guided neural network outperforms the state-of-the-art method for HAR in smart homes across multiple datasets and by large margins. These results are promising because they push HAR for smart homes closer to real-world applications.

Towards a data-driven adaptive anomaly detection system for human activity

• An adaptive system is proposed for abnormality detection in human activities. • The data-driven system adapts to changes in human behavioural routine. • The approach is based on data ageing and data dissimilarity forgetting factors. • The forgetting factor features allow the system to discard old behavioural patterns. Research in the field of ambient intelligence allows for the utilisation of different computational models for human activity recognition and abnormality detection to promote independent living and to improve the quality of life for the increasing ageing population. The existing monitoring systems are not adaptive to the overly changing human behavioural routine leading to a high rate of false predictions. An adaptive system pipeline is proposed in this paper for adapting to changes in human behaviour based on data ageing and data dissimilarity forgetting factors. The forgetting factor feature allows adaptation of the model to the current routines of an individual while forgetting outdated behavioural patterns. The data ageing forgetting factor discard old behavioural routine based on the age of the activity data while in the data dissimilarity approach, this is achieved by measuring the similarity of the activity data. Behaviour modelling is achieved using an ensemble of novelty detection models termed as Consensus Novelty Detection Ensemble consisting of One-Class Support Vector Machine, Local Outlier Factor, Robust Covariance Estimation and Isolation Forest. The proposed approach is data-driven and environment-invariant, making it feasible for deployment in heterogeneous environments. A comparative analysis carried out with other abnormality detection models for human activities across two datasets shows that the proposed approach achieved better results.

Grand challenges for ambient intelligence and implications for design contexts and smart societies

This paper highlights selected grand challenges that concern especially the social and the design dimensions of research and development in Ambient Intelligence (AmI) and Smart Environments (SmE). Due to the increasing deployment and usage of ‘smart’ technologies determining a wide range of everyday life activities, there is an urgent need to reconsider their societal implications and how to address these implications with appropriate design methods. The paper presents four perspectives on the subject grounded in different approaches. First, introducing and reflecting on the implications of the ‘smart-everything’ paradigm, the resulting design trade-offs and their application to smart cities. Second, discussing the potential of non-verbal communication for informing the design of spatial interfaces for AmI design practices. Third, reflecting on the role of new data categories such as ‘future data’ and the role of uncertainty and their implications for the next generation of AmI environments. Finally, debating the merits and shortfalls of the world’s largest professional engineering community effort to craft a global standards body on ethically aligned design for autonomous and intelligent systems. The paper benefits from taking different perspectives on common issues, provides commonalities and relationships between them and provides anchor points for important challenges in the field of ambient intelligence.

Hierarchical classification method based on selective learning of slacked hierarchy for activity recognition systems

Abstract Physical activity recognition using wearable sensors has gained significant interest from researchers working in the field of ambient intelligence and human behavior analysis. The problem of multi-class classification is an important issue in the applications which naturally has more than two classes. A well-known strategy to convert a multi-class classification problem into binary sub-problems is the error-correcting output coding (ECOC) method. Since existing methods use a single classifier with ECOC without considering the dependency among multiple classifiers, it often fails to generalize the performance and parameters in a real-life application, where different numbers of devices, sensors and sampling rates are used. To address this problem, we propose a unique hierarchical classification model based on the combination of two base binary classifiers using selective learning of slacked hierarchy and integrating the training of binary classifiers into a unified objective function. Our method maps the multi-class classification problem to multi-level classification. A multi-tier voting scheme has been introduced to provide a final classification label at each level of the solicited model. The proposed method is evaluated on two publicly available datasets and compared with independent base classifiers. Furthermore, it has also been tested on real-life sensor readings for 3 different subjects to recognize four activities i.e. Walking, Standing, Jogging and Sitting. The presented method uses same hierarchical levels and parameters to achieve better performance on all three datasets having different number of devices, sensors and sampling rates. The average accuracies on publicly available dataset and real-life sensor readings were recorded to be 95% and 85%, respectively. The experimental results validate the effectiveness and generality of the proposed method in terms of performance and parameters.

A Network of Stationary Sensors and Mobile Robots for Distributed Ambient Intelligence

In the last few decades, sensor networks have received significant attention in the field of ambient intelligence (AmI) for surveillance and assisted living applications, as they provide a powerful tool to capture relevant information about environments and human activities. Mobile robots hold promise for enhancing the potential of sensor networks toward the development of intelligent systems that are able not only to detect events but also to actively intervene in the environment accordingly. This article presents the Distributed Ambient Intelligence Architecture that aims at integrating multisensor robotic platforms with wireless sensor networks. Based on the robot operating system, it provides a flexible and scalable software infrastructure extendible to different AmI scenarios. This article describes the proposed architecture and presents experimental tests, showing the feasibility of the system in the context of ambient assisted living.

Ambient Intelligence: Vision, research, and life

This paper evaluates the progress in the field of Ambient Intelligence since its early days. It discusses the extent to which earlier visions of Ambient Intelligence have materialized arguing that to a large extent we already live in an era of Ambient Intelligence, despite that ‘true intelligence’ appears to remain elusive. We examine the extent to which this is so, focusing on shifting mind-sets and ambitions for researchers in the last years. We discuss the various reincarnations of the vision from a historical and societal perspective, together with some of the critiques leveled against it. With many research challenges still remaining open, we discuss how Ambient Intelligence research can be aligned to contemporary societal needs. Particularly important is to create mechanisms that will enable people to interact and benefit from the envisioned technological infrastructure avoiding some of the dangers that embedding computational intelligence in our physical environment brings about. Observing the diversification of research to address different application areas, we note how the impetus towards Ambient Intelligence needs to be renewed with a uniting challenge that transcends the development of technological infrastructure.

Electromyographic Patterns during Golf Swing: Activation Sequence Profiling and Prediction of Shot Effectiveness.

This study analyzes muscle activity, recorded in an eight-channel electromyographic (EMG) signal stream, during the golf swing using a 7-iron club and exploits information extracted from EMG dynamics to predict the success of the resulting shot. Muscles of the arm and shoulder on both the left and right sides, namely flexor carpi radialis, extensor digitorum communis, rhomboideus and trapezius, are considered for 15 golf players (∼5 shots each). The method using Gaussian filtering is outlined for EMG onset time estimation in each channel and activation sequence profiling. Shots of each player revealed a persistent pattern of muscle activation. Profiles were plotted and insights with respect to player effectiveness were provided. Inspection of EMG dynamics revealed a pair of highest peaks in each channel as the hallmark of golf swing, and a custom application of peak detection for automatic extraction of swing segment was introduced. Various EMG features, encompassing 22 feature sets, were constructed. Feature sets were used individually and also in decision-level fusion for the prediction of shot effectiveness. The prediction of the target attribute, such as club head speed or ball carry distance, was investigated using random forest as the learner in detection and regression tasks. Detection evaluates the personal effectiveness of a shot with respect to the player-specific average, whereas regression estimates the value of target attribute, using EMG features as predictors. Fusion after decision optimization provided the best results: the equal error rate in detection was 24.3% for the speed and 31.7% for the distance; the mean absolute percentage error in regression was 3.2% for the speed and 6.4% for the distance. Proposed EMG feature sets were found to be useful, especially when used in combination. Rankings of feature sets indicated statistics for muscle activity in both the left and right body sides, correlation-based analysis of EMG dynamics and features derived from the properties of two highest peaks as important predictors of personal shot effectiveness. Activation sequence profiles helped in analyzing muscle orchestration during golf shot, exposing a specific avalanche pattern, but data from more players are needed for stronger conclusions. Results demonstrate that information arising from an EMG signal stream is useful for predicting golf shot success, in terms of club head speed and ball carry distance, with acceptable accuracy. Surface EMG data, collected with a goal to automatically evaluate golf player’s performance, enables wearable computing in the field of ambient intelligence and has potential to enhance exercising of a long carry distance drive.

Improving the Utilization of AAL Devices through Semantic Web Technologies and Web of Things Concepts

The software-driven assistance of elderly and impaired people in their every day life requires the integration of heterogeneous devices and assistive services. The lack of interoperability hampers the integration and usage of these devices and services. Furthermore, one objective in the field of Ambient Intelligence and Human Machine Interaction is the adaption of assistive services to the user 5. This is deeply rooted in the idea that the usage of ambient technologies still means a barrier for elderly and impaired people and prevents them to trust and use these technologies. But as these technologies aim at supporting and not restricting elderly and impaired people, a methodology is needed to overcome the barrier of using ambient assistive technologies. In this paper, we present an approach to a) enable a simplified integration of heterogeneous devices and services by the Web of Things recommendation and b) a lightweight AAL ontology together with a rule-based engine called Sherlock in order to demonstrate the advantages in representing and linking user intentions to device capabilities using Semantic Web technologies. Moreover, we show how this semantic representation in the context-aware Sherlock engine is used to derive user intentions by recognising device events. The practical applicability of the presented approach is demonstrated in a real-world use case from the AICASys project1.1See: http://www.mtidw.de/ueberblick-bekanntmachungen/ALS/aicasys

Ambient intelligence and ergonomics in Asia

Ambient intelligence (AmI) is one of the most important advances in computer technology in the mobile era. By being aware and applying knowledge of ergonomics (human factors), AmI system users will be in better shape and health, become more effective and find using AmI system a more pleasant and enjoyable experience. This special issue contains selected papers from the 1st international conference on ambient intelligence and ergonomics in Asia (AmIE 2013), which was held at Taichung city, Taiwan, July 3–5, 2013. The conference attracted a large number of scientific papers that contributed to the state-of-the-art in the fields of ambient intelligence and ergonomics. After a strict review, five articles from researchers around the world were finally accepted. To determine an understandable algorithm that could be used by designers to create novel concept designs, M.T. Wang and C.C. Yang selected the popular motor scooter as a sample product, and used the most distinctive front handle cover as a design target. Their method included three phases: preparation, construction of conceptual creativity, and semantic analysis. After comparing random idea sketches with designs available on the market, they observed that certain concept designs obtained using this method were exceptionally innovative, and could be easily redesigned for an actual product. A.M. Otebolaku and M.T. Andrade investigated context-aware recommendation techniques for implicit delivery of contextually relevant online media items. The proposed recommendation service identifies a user’s dynamic contextual situation from the device’s built-in sensors, and uses case-based reasoning to determine the user’s current contextual preferences. The effectiveness of the proposed recommendation service was evaluated with a case study. Gripping and pinching are frequently used hand strength in various occupational activities and in clinical evaluation of the hand. Therefore, formulating grip and pinch prediction models with easily obtainable personal parameters will help facilitate the design and evaluation of workplace environments or facilitate the hand impairment or progression assessments. P.C. Sung, C.C. Hsu, C.L. Lee, Y.-S.P. Chiu, and H.L. Chen developed maximum voluntary contraction (MVC) grip and key pinch strength prediction models using regression method and artificial neural networks (ANN). An ANN is a prevalent humanized computing method that imitates the central nervous system of a human. To estimate the cycle time of a job in a wafer fabrication factory, H.C. Wu and T. Chen proposed a joint use of a classification and regression tree (CART) and back propagation network (BPN). In their method, a BPN is constructed to estimate the cycle times of jobs of a branch. A real case was used to evaluate the effectiveness of the proposed methodology. C.T. Tseng, Y.L. Lee, and C.C. Chou designed an AmIbased decision support system that combined an electromagnetism-like mechanism (EM) and sensory data to aid human operators in making decisions regarding the T. Chen (&) Department of Industrial Engineering and Systems Management, Feng Chia University, Taichung, Taiwan e-mail: tcchen@fcu.edu.tw

Ambient Activity Recognition in Smart Environments for Cognitive Assistance

In this paper, the authors investigate the challenging key issues that emerge from research in the field of ambient intelligence in smart environments, under the context of activity recognition. The authors clearly describe the specific functional needs inherent in cognitive assistance for effective activity recognition, and then the authors present the fundamental research that addresses this problem in such a context. This paper is more of a survey and an analysis of existing works that have been studied for potential integration into our laboratories, rather than a focused evaluation report. The authors’ objective is to identify gaps in the capabilities of current techniques and to suggest the most productive lines of research to address this complex issue. As such, the contribution is of both theoretical and practical significance.

A middleware to integrate robots, simple devices and everyday objects into an ambient ecology

The fields of ambient intelligence, distributed robotics and wireless sensor networks are converging toward a common vision, in which ubiquitous sensing and acting devices cooperate to provide useful services in the home. These devices can range from sophisticated mobile robots to simple sensor nodes and even simpler tagged everyday objects. In this vision, a milkbox left on the table after the user has left the home could ask the service of a mobile robot to be placed back in the refrigerator. A missing ingredient to realize this vision is a mechanism that enables the communication and interoperation among such highly heterogeneous entities. In this paper, we propose such a mechanism in the form of a middleware able to integrate robots, tiny devices and augmented everyday objects into one and the same system. The key moves to cope with heterogeneity are: the definition of a tiny, compatible version of the middleware, that can run on small devices; and the concept of object proxy, used to make everyday object accessible within the middleware. We describe the concepts and implementation of our middleware, and show a number of experiments that illustrate its performance.

Ambient e-Learning: a metacognitive approach

In the last years, the importance of context-aware software applications able to provide personalized and adaptive services has been growing. Moreover, new innovative technologies make it possible to embed context-aware applications into everyday life objects in order to implement seamless user experiences. The research in the field of Ambient Intelligence is focused on these technologies. In particular, Ambient Learning denotes the presence of new ICTs embedded into the environment which leads to advanced e-Learning scenarios. The literature provides a set of requirements, which characterize the Ambient Learning processes that are required to be permanent, goal-directed, interactive, embedded in dayly life situations, personalized and context-aware. Both frameworks and technological mappings are provided to drive the definition of ambient e-Learning systems, nevertheless Self Regulated Learning is partially addressed, and it is not clear how these frameworks can enable Self Regulatory Processes to foster life-long learning. The goal of the present work is to address the aforementioned criticisms, providing a new framework for Ambient Learning that may also leverage on collective knowledge and objective-driven learning aspects to be exploited in synergy with the pillars of existing approaches.

Responsive environments: User experiences for ambient intelligence

In this paper we review the emerging field of responsive environments as an ambient intelligence system with a focus on user experience. We see responsive environments as a combination between the scientific developments that resulted in Ambient Intelligence systems and the aesthetic motivation behind interactive art installations. This next development in Information and Communication Technologies transcends the computer screen and the localised interface towards ambient accessibility and experience. Inspired by different authors in the field of Ambient Intelligence, we redefine Responsive environments as physical spaces, such as city squares, public halls, etc, that are enhanced with the use of technology and media. The concept of responsive environments is defined in relation to ambient intelligence highlighting system-centred approach based on which it has been developed, and user-centred approach that we believe is the direction to follow. The infrastructure and technology required, notably the necessary sensors and actuators are reviewed in later sections. The user interaction is analysed with a series of examples. The different approaches to the interfaces are contributing to a better understanding of the challenges of selecting the right interaction principle. Furthermore, context awareness and user experience have an important contribution to the success of a responsive environment. An environment where the aesthetic of interaction, user engagement, access, embodiment and intimacy are also to be taken into account in the design and specifications phases.

Ambient intelligence and wearable computing: Sensors on the body, in the home, and beyond

Ambient intelligence has a history of focusing on technologies that are integrated into a person's environment. However, ambient intelligence can be found on a person's body as well. In this thematic issue we examine the role of wearable computing in the field of ambient intelligence. In this article we provide an overview of the field of wearable computing and discuss its relationship to the fields of smart environments and ambient intelligence. In addition, we introduce the papers presented in the thematic issue highlighting a number of research projects which are defining the state of the art in wearable computing and ambient intelligence.

Embedded Technology: Current and Future Practices for Increasing Accessibility for All Students

This article highlights the technological solutions available for increasing the accessibility of educational materials for students with disabilities. Special attention is paid to design features for supporting the needs of students with disabilities in computer-based tests; however, many of the same technologies can be applied to instructional environments. Current technologies such as assistive devices are discussed as well as emerging strategies for embedding flexibility into learning and testing interfaces. In addition, advances in the field of ambient intelligence are described as they might apply to educational environments.