Search

Recognizing activities of daily living is useful for ambient assisted living. In this regard, the use of wearable cameras is a promising technology. In this paper, we propose a novel approach for recognizing activities of daily living using egocentric viewpoint video clips. First, in every frame, the appearing objects are detected and labelled depending if they are being used or not by the subject. Later, the video clip is divided into spatiotemporal bins created with an object-centric cut. Finally, a support vector machine classifier is computed using a spatiotemporal flexible kernel between video clips. The validity of the proposed method has been proved by conducting experiments in the ADL dataset. Results confirm the suitability of using the space-time location of objects as information for the classification of activities using an egocentric viewpoint.

Wearable and Internet of Things (IoT) devices have the potential to improve lifestyle, personalize receiving treatments or introduce assisted living for elderly people. However, service delivery depends on maintaining and troubleshooting device connectivity to smartphones, where user engagement and technology proficiency represent a possible barrier that prevents a wider adoption, especially in the elderly and disabled population. Low-cost and low-power wearable and IoT devices face challenges when operating out of range of known home networks or pared devices. We propose an architecture and methods to provide seamless connectivity (Se-Co) between devices and wireless networks while maintaining low-power, low-cost and standards compatibility. Through Se-Co, the devices connect without user interaction both in home and in unknown roaming networks while maintaining anonymity, privacy and security. Roaming networks approve data limited connectivity to unknown devices that are able to provide a valid anonymized certificate of compliance and no harm through a home provider. Se-Co enables shifting data processing, such as pattern processing using artificial intelligence, from a wearable device or smartphone towards the cloud. The proposed Se-Co architecture could provide solutions to increase usability of wearable devices and improve their wider adoption, while keeping low the costs of devices, development and services.

Discovering routine behavior patterns of the solitary elderly can provide the elderly with better assisted living services. This paper presents a method based on time, path and area association restriction, to calculate the distance between the solitary elderly daily behavior patterns and then discover the routine behavior patterns.

Internet of Things (IoT) devices are rapidly penetrating e-health and assisted living domains, and an increasing proportion among them goes on the account of computationally-weak devices, where security and privacy provisioning alone are demanding tasks, not to mention grouping proofs. This paper, therefore, gives an extensive analysis of such proofs and states lessons learnt to avoid possible pitfalls in future designs. It sticks with prudent engineering techniques in this field and deploys in a novel way the so called non-deterministic principle to provide not only grouping proofs, but (among other) also privacy. The developed solution is analyzed by means of a tangible metric and it is shown to be lightweight, and formally for security.

Activity recognition(AR) is a popular subject of research in the recent past. Recognition of activities performed by human beings, enables the addressing of challenges posed by many real-world applications such as health monitoring, providing security etc. Segmentation plays a vital role in AR. This paper evaluates the efficiency of Area-Based Segmentation using different performance measures. Area-Based segmentation was proposed in our earlier research work. The evaluation of the Area-Based segmentation technique is conducted on four real world datasets viz. Aruba17, Shib010, HH102, and HH113 comprising of data pertaining to an individual, living in the test bed home. Machine learning classifiers, SVM-R, SVM-P, NB and KNN are adopted to validate the performance of Area-Based segmentation. Amongst the four chosen classification algorithms SVM-R exhibits better in all the four datasets. Area-Based segmentation recognise the four test bed activities with accuracies of 0.74, 0.98, 0.66, and 0.99 respectively. The results reveal that Area based segmentation can efficiently segment sensor data stream which aids in accurate recognition of smart home activities.

Human activity recognition (HAR) from sensor readings has proved to be an effective approach in pervasive computing for smart healthcare. Recent approaches in ambient assisted living (AAL) within the home or community setting offers people the prospect of independent care and improved quality of living. However, most of the available AAL systems are limited by several factors including the system complexity and computational cost. In this paper, a simple, the novel ambient HAR framework using the multivariate Gaussian is proposed. The classification framework augments prior information from passive RFID tags to obtain more detailed activity profiling. The proposed algorithm based on the multivariate Gaussian via maximum likelihood estimation is used to learn the features of the human activity model. The twelve sequential and concurrent experimental evaluations are conducted in a mock apartment environment. The sampled activities are predicted using a new dataset of the same activity and high prediction accuracy established. The proposed framework suits well for the single and multi-dwelling environment and offers pervasive sensing environment for elderly, disabled, and carers.

International audience; Human action recognition in videos is an important issue in computer vision. We propose an approach based on the integration of partial or global human body segmentation in the classification process to deal with partial movements and immobility. Experimentation on UCF101 public dataset output competitive recognition accuracy related state of the art.

Human activity recognition emerges as a rising field of investigation in achieving brilliant medicinal services observing and benefits for ambient assisted living (AAL). By using the signs amassed from different sensors inbuilt in cell phones, the HAR can perceive the exercises of subjects. For our work, seven unique kinds of exercises (online available dataset) are utilized and Neighborhood Component Analysis (NCA) is utilized to choose the significant features from the huge dataset. In this way, the computational cost is reduced. The grouping models, for example, Random Forest and K-Nearest Neighbor are utilized to distinguish the seven distinctive human exercises. The activities are classified using KNN and Random Forest classifiers with different neighbors and results are tabulated.

Smart homes and assisted living facilities are better able to improve the lives of residents if they can perform intent recognition. If the system understands the goals of residents it can select appropriate assistive actions in advance, for example. The increasing availability of cheap sensors and smart devices provides fine grained data about residents and the ambient environment. Making full use of this data for intent recognition requires both relational and probabilistic reasoning. In this paper we present a framework for performing combined logical and probabilistic reasoning about intentions using data from simple switch sensors. We employ ProbLog, a probabilistic extension to Prolog, to infer the most probable intention given observations of the actions of the agent and sensor readings of important aspects of the environment.

With advances in machine learning, the evaluation and analysis of human behavior continue to attract large number of researchers around the globe. In this paper, we furnish an extensive overview of ways to identifying, analyzing and assessing human behavior, taking into account various behavioral characteristics. Most promising attributes and recognition techniques for vision and sensor-based approaches have been detailed. Most prominently used datasets for both vision- and sensor-based approaches have also been studied, keeping in mind the nature, source and applications of the same in the field of human behavior and activity detection. The study indicates that sensor-based approaches tend to have an upper hand because of the privacy breach caused by vision-based approaches, which accounts for the evolving usage of sensor-based monitoring for real-time behavior detection. Various other deep learning methods and their applications in the field of behavioral recognition have also been stated.

There is a considerable demand for human activity recognition techniques in the area of human perception and also encompasses many other purposes like healthcare monitoring, assisted living for elders, and intelligent video surveillance. There are different approaches to machine learning that have been adapted for the purpose of activity recognition. But these techniques depend heavily on hand-crafted feature extraction which is unable to perform well when dealing with complex scenarios. Deep learning techniques have great potential for human activity recognition. In this paper, a neural network (NN) based approach for classification and evaluation of human activities has been explored. In this method, a convolutional neural network (CNN) is put together with long short-term memory (LSTM). The dataset experimented in this system is the classic Human Activity Recognition (HAR) dataset for classifying the six human activities, viz., walking, walking-upstairs, walking-downstairs, sitting, standing, and laying. Results show that the proposed model is very efficient for recognizing human activity.

This research discusses about a way to identify appliances in the smart space using artificial neural network (ANN). Smart appliances identification technique can solve a few specific problems for AAL like reducing the implementation complexity in the application level, finding the anomaly in the early stages. We have proposed an ANN-based approach to identify the appliances using smart space data. The smart appliances identification process is initiated by collecting data related to power consumption of smart appliances which are used for AAL services. An individual template along ANN-based approach is proposed to improve the accuracy. Through the interaction of the template, the proposed system has the ability to improve its identification by labeling the unlabeled input data. An ANN-based feedforward algorithm multilayer perceptron (MLP) is used for identifying the appliances. The output from MLP is observed whether the data pattern is a new class or not. The template provides the possible solution based on features and data. The proposed model has achieved identification performance with an average of 93% accuracy.

Implantable Medical Devices (IMDs) play a very critical role in both medical and non-medical fields. Hence, protecting the security and privacy of these devices is among the highest priorities, as failing to do so would jeopardize the life of the patient. One way of maintaining the safety of IMDs is to design an authentication protocol between the IMDs and the external devices. With this regard, although several authentication schemes exist, a significant number of these protocols did not include formal verification to prove their secrecy against known attacks. One such authentication scheme is the He and Zeadally protocol for IMD-Enabled Ambient Assisted Living System. Accordingly, in this paper, we analyzed the security of this protocol by using formal verification methods -BAN-Logic and AVISPA. As a result, despite the protocol's strong qualities, we found that the protocol is insecure and fell short of other essential such as emergency authentication and key-update procedures.

International audience; Recent improvements in connected tools and learning algorithms allow new opportunities in the field of Ambient Assisted Living (AAL). However, smart home inhabitant's life habits are often required to obtain adequate results for energy management, security, Health at Home (HaH), and numerous other applications. In this paper, a model for life routines representation and algorithms for its generation is introduced. Study on the state of the art exposes that activity ordering and duration are key features of human behavior. Consequently, the presented approach focuses on a higher level of semantic by observing activities performed by the inhabitant rather than the sensor logs, which allow for better understanding of his comportment and universality of the model for multiple aims. Stochastic Time Automata (STA) is proposed as it adequately models activity ordering with probability associated to edges and activity duration through probability distribution associated to location delay. Presented approach does not require specific equipment besides sensors required for activity recognition and is versatile enough to be used in various applications. A case study highlights the relevancy of the chosen features and demonstrates that the proposed model is efficient to depict and understand inhabitants' life habits.

Ambient Assisted Living (AAL) is a research area, where the focus lies on enabling people with any kind of impairment to stay independent in their own home for as long as possible...

Active and Assisted Living (AAL) systems have been developed over the last decade to improve the quality of life, independence and well-being of the elderly. In the future, there will be an increased need to embed context sensitive AAL systems in an existing smart environment, equipped with built-in sensors, wearable sensors or smart devices. Especially in the health sector, AAL system architectures show similarities with application architectures in the field of Smart Homes or Smart Cities. In all these domains the use of context-based middleware is essential. Internet of Things (IoT) platforms, which are used particularly in the area of Smart Homes or Smart Cities, deal with a wider range of data than AAL middleware, for which there is no generally accepted solution yet. The aim of this paper is to present a flexible, model-based coupling of context-based middleware, which is widely accepted especially in the field of Smart Cities, with standard interfaces for use in the AAL area. In this paper we use the project AYUDO as an AAL use case and FIWARE, an open IoT middleware platform, as a context-based middleware candidate for this coupling.

Research projects in the field of Active and Assisted Living (AAL) are dedicated to enhance quality of life and independent living of the elderly by providing assistive technical and non-technical solutions. With the progress of the Age of Internet of Things (IoT) and digitalization, the implemented applications are also often focused on the integration of IoT technologies. As the field of AAL is characterized by a high amount of interdisciplinarity, there arises the need of integrating various proprietary solutions to one comprehensive application. This can be achieved by using dedicated IoT middleware solutions which provide the possibility to integrate information from different systems, users and devices into one comprehensive application. The aim of this paper is to show an overview on the process of interdisciplinary research in the field of AAL and to demonstrate the arising requirements for IoT middleware solutions in the given context. Furthermore, selected middleware frameworks and platforms are presented and the implementation of one selected solution is described for a given use case from the Smart VitAALity project.

The application of blockchains techniques in the Internet of Things (IoT) is gaining much attention with new solutions proposed in diverse areas of the IoT. Conventionally IoT systems are designed to follow the centralised paradigm where security and privacy control is vested on a “trusted” third-party. This design leaves the user at the mercy of a sovereign broker and in addition, susceptible to several attacks. The implicit trust and the inferred reliability of centralised systems have been challenged recently following several privacy violations and personal data breaches. Consequently, there is a call for more secure decentralised systems that allows for finer control of user privacy while providing secure communication. Propitiously, the blockchain holds much promise and may provide the necessary framework for the design of a secure IoT system that guarantees fine-grained user privacy in a trustless manner. In this paper, we propose a holistic blockchain-based decentralised model for Ambient Assisted Living (AAL) environment. The nodes in our proposed model utilize smart contracts to define interaction rules while working collaboratively to contribute storage and computing resources. Based on the blockchain technique, our proposed model promotes trustless interaction and enhanced user’s privacy through the blockchain - Interplanetary File System (IPFS) alliance. The proposed model also addresses the shortfall of storage constraints exhibited in many IoT systems.

Ambient assisted living plays a vital role in assisting the elderly people. There will be a significant amount of rise in the population of the elderly people in the future. Elderly people must need an assistance to enrich their lifestyle. To provide such an opportunity, human activity recognition (HAR) will be of great use. Monitoring day to day activities are essential to assist independent elders. In this paper, we have collected a dataset of 5 daily activities using accelerometer and gyroscope from elders and used data mining techniques such as KNN and Random forest for classification. We have achieved an accuracy of 88 % using KNN and 89% using random forest. The results obtained are favorable in recognizing the activities to the greater extent.

As the number of older adults increases worldwide, new paradigms for indoor activity monitoring are required to keep people living at home independently longer. Radar-based human activity recognition has been identified as a sensing modality of choice because it is privacy-preserving and does not require end-users compliance or manipulation. In this paper, we explore the robustness of machine learning algorithms for human activity recognition using six different activities from the University of Glasgow dataset recorded with an FMCW radar. The raw radar data is pre-processed and represented using four different domains, namely, range-time, range-Doppler amplitude and phase diagrams, and Cadence Velocity Diagram. From those, salient features can be extracted and classified using Support Vector Machine, Stacked AutoEncoder, and Convolutional Neural Networks. The fusion of handcrafted features and features from CNN is applied to get the best scheme of classification with over 96% accuracy.

This paper proposes “An Emergency Situation Detection System for Ambient Assisted Living (AAL)”, to support elderly people and patients with chronic conditions and potential health-related emergencies to live independently. It implements an Internet of Things (IoT) network that continuously monitors the health conditions of these people. The network includes mobile phones, to transmit the data generated by the IoT sensors to the cloud server. Especially, the paper proposes the 3 rd party unknown mobile relays instead of dedicated gateways as opposed to many existing solutions for IoT healthcare applications. The wireless communication technology used to provide the connectivity between the sensor nodes and mobile relays is Bluetooth Low Energy (BLE). To establish a secure end-to-end connectivity between low power IoT sensor nodes and cloud servers, the paper proposes several techniques. After the medical data transmission to the cloud server, it is responsible for emergency detection and alert generation accordingly. The type of emergency is not limited to a specific health issue, but new emergency situations can be defined and added to the proposed system. Ultimately, the interested parties such as family members, caretakers and doctors receive these alerts. The development of a prototype of the system as a part of the work using commercial off-the-shelf devices verifies the validity of the proposing system and evaluates the performance advantage over the existing systems.

A Convolutional Neural Network (CNN) is sometimes confronted with objects of changing appearance ( new instances) that exceed its generalization capability. This requires the CNN to incorporate new knowledge, i.e., to learn incrementally. In this paper, we are concerned with this problem in the context of assisted living. We propose using the feature space that results from the training dataset to automatically label problematic images that could not be properly recognized by the CNN. The idea is to exploit the extra information in the feature space for a semi-supervised labeling and to employ problematic images to improve the CNN's classification model. Among other benefits, the resulting semi-supervised incremental learning process allows improving the classification accuracy of new instances by 40% as illustrated by extensive experiments., Accepted in VISAPP 2020

In the recent years, the advancement of technology, the constantly aging population and the developments in medicine have resulted in the creation of numerous ambient assisted living systems. Most of these systems consist of a variety of sensors that provide information about the health condition of patients, their activities and also create alerts in case of harmful events. Successfully combining and utilizing all the multimodal information is an important research topic. The current paper compares model-based and class-based fusion, in order to recognize activities by combining data from multiple sensors or sensors of different body placements. More specifically, we tested the performance of three fusion methods; weighted accuracy, averaging and a recently introduced detection rate based fusion method. Weighted accuracy and the detection rate based fusion achieved the best performance in most of the experiments.

We explain how a solution for data privacy, and specifically for cognitive security, can be enforced and guaranteed using blockchain technology in SAAL (Smart Ambient Assisted Living) environments. Using our proposal the access to a patient’s clinical process is secure for the adequate interested and authorized parties while resist tampering and ransomware attacks that have recently plagued the HIS (Hospital Information Systems) in various countries.Journal of the Institute of Engineering, 2017, 13(1): 37-47

Various surveillance systems capture signs of human activities of daily living (ADLs) and store multimodal information as time line behavioral records. In this paper, we present a novel approach to the analysis of a behavioral record used in a surveillance system designed for use in elderly smart homes. The description of a subject's activity is first decomposed into elementary poses - easily detectable by dedicated intelligent sensors - and represented by the share coefficients. Then, the activity is represented in the form of an attributed graph, where nodes correspond to elementary poses. As share coefficients of poses are expressed as attributes assigned to graph nodes, their change corresponding to a subject's action is represented by flow in graph edges. The behavioral record is thus a time series of graphs, which tiny size facilitates storage and management of long-term monitoring results. At the system learning stage, the contribution of elementary poses is accumulated, discretized and probability-ordered leading to a finite list representing the possible transitions between states. Such a list is independently built for each room in the supervised residence, and employed for assessment of the current action in the context of subject's habits and a room purpose. The proposed format of a behavioral record, applied to an adaptive surveillance system, is particularly advantageous for representing new activities not known at the setup stage, for providing a quantitative measure of transitions between poses and for expressing the difference between a predicted and actual action in a numerical way.

Several pathologies can alter the way people walk, i.e., their gait. Gait analysis can be used to detect such alterations and, therefore, help diagnose certain pathologies or assess people’s health and recovery. Simple vision-based systems have a considerable potential in this area, as they allow the capture of gait in unconstrained environments, such as at home or in a clinic, while the required computations can be done remotely. State-of-the-art vision-based systems for gait analysis use deep learning strategies, thus requiring a large amount of data for training. However, to the best of our knowledge, the largest publicly available pathological gait dataset contains only 10 subjects, simulating 5 types of gait. This paper presents a new dataset, GAIT-IT, captured from 21 subjects simulating 5 types of gait, at 2 severity levels. The dataset is recorded in a professional studio, making the sequences free of background camouflage, variations in illumination and other visual artifacts. The dataset is used to train a novel automatic gait analysis system. Compared to the state-of-the-art, the proposed system achieves a drastic reduction in the number of trainable parameters, memory requirements and execution times, while the classification accuracy is on par with the state-of-the-art. Recognizing the importance of remote healthcare, the proposed automatic gait analysis system is integrated with a prototype web application. This prototype is presently hosted in a private network, and after further tests and development it will allow people to upload a video of them walking and execute a web service that classifies their gait. The web application has a user-friendly interface usable by healthcare professionals or by laypersons. The application also makes an association between the identified type of gait and potential gait pathologies that exhibit the identified characteristics. info:eu-repo/semantics/publishedVersion

Human activity recognition has the potential to become a real enabler for ambient assisted living technologies. Research on this area demands the execution of complex experiments involving humans interacting with intelligent environments in order to generate meaningful datasets, both for development and validation. Running such experiments is generally expensive and troublesome, slowing down the research process. This paper presents an agent-based simulator for emulating human activities within intelligent environments: MASSHA. Specifically, MASSHA models the behaviour of the occupants of a sensorised environment from a single-user and multiple-user point of view. The accuracy of MASSHA is tested through a sound validation methodology, providing examples of application with three real human activity datasets and comparing these to the activity datasets produced by the simulator. Results show that MASSHA can reproduce behaviour patterns that are similar to those registered in the real datasets, achieving an overall accuracy of 93.52% and 88.10% in frequency and 98.27% and 99.09% in duration for the single-user scenario datasets; and a 99.3% and 88.25% in terms of frequency and duration for the multiple-user scenario.

Internet of Things (IoT) is the logical further development of today's Internet, enabling a huge amount of devices to communicate, compute, sense and act. IoT sensors placed in Ambient Assisted Living (AAL) environments, enable the context awareness and allow the support of the elderly in their daily routines, ultimately allowing an independent and safe lifestyle. The vast amount of data that are generated and exchanged between the IoT nodes require innovative context modeling approaches that go beyond currently used models. Current paper presents and evaluates an open interoperable platform architecture in order to utilize the technical characteristics of IoT and handle the large amount of generated data, as a solution to the technical requirements of AAL applications.

This paper presents DemaWare2, an Ambient Assisted Living framework to support the care of people with dementia. The framework integrates various sensor modalities, such as ambient, wearable, offline and cloud-based, together with sophisticated, interdisciplinary methods including image, audio and semantic analysis. Fine-grained, atomic events, such as object manipulation, are aggregated into complex activities through semantic fusion. Applications tailored to monitoring dementia symptoms support clinicians to drive effective, timely interventions and evaluate their outcomes. The framework was evaluated for its robustness, reliability and clinical value in real-world lab trials and home installations.

The success of providing smart healthcare services in ambient assisted living (AAL) largely depends on an effective prediction of situations in the environment. Situation awareness in AAL is to determine the environment smartness by perceiving information related to the surroundings and human behavioral changes. In AAL environment, there are plenty of ways to collect data about its inhabitants, such as through cameras, microphones, and other sensors. The collected data are complicated enough to go for an efficient processing in perceiving the situation. This paper gives an overview of the existing research results in multimodal data analysis in AAL environment to improve the living environment of the seniors, and it attempts to bring efficiency in complex event processing for real-time situational awareness. This paper thus considers multimodal sensing for detection of current situations as well as to predict future situations using decision-tree and association analysis algorithms. To illustrate the proposed approach, we consider elderly activity recognition in the AAL environment.

The IEEE Sensors Applications Symposium (SAS) 2019 was the 14th edition of this conference. SAS, one of the flagship conferences of the IEEE Instrumentation and Measurement Society, brings together sensor developers, innovators, and users, providing a forum for the exploration of new applications in sensor technology. SAS 2019 was very successful thanks to a large number of high-quality articles and the numerous participants who attended the conference, one-third of them were students. A wide range of subject areas was covered during SAS 2019, such as sensor data fusion, energy harvesting for sensors, biomedical sensors and systems, industrial applications, novel sensing technologies, smart building, and smart agriculture. In addition, hot topics, such as assisted living for the aging population, flexible and wearable sensors, sensors for automotive applications, and smart IoT sensing and networking, were also addressed.

Advances in Bluetooth technology focusing low energy has brought into action the Bluetooth Smart for wearable devices. Enhancement in Internet of Things due to wearable devices has led to the concept of Ambient Assisted Living (AAL). AAL is mainly devised to assist age old people to carry on daily activities including healthcare without dependencies. A survey to show the contribution of Bluetooth Low Energy (BLE) technology in wearable devices is the main focus of this paper. This paper also aims to give a good overview of BLE and BLE Beacons as one of the successful technologies focusing wearable healthcare applications and its pros and cons compared to other existing wireless technologies. Different applications of BLE included can be adapted to our healthcare perspective for smart health management.

A significant challenge to provide services to the inhabitant in a smart environment resides in the effective implementation of models. Most of the proposed models are conceptual and lacks practical consideration. Human activity recognition is one of the most challenging tasks to offer the solution for ambient assisted living. In this work, we explore a time series classification problem - human activity recognition. Total of nine machine learning and deep learning algorithms implemented and evaluated using the same dataset. The results are analyzed using different parameters. This paper aims at providing help to select a practical machine learning approach for activity recognition process in ambient assisted living systems. The comparative analysis shows that in deep learning, Long Short-Term Memory (LSTM) network performed best with a classification accuracy of 92%. In machine learning, Logistic Regression and Gradient Boosting came out with a classification accuracy of greater than 90%, and others came out the worst with classification accuracy less than 90%.

In this paper we present a real-time demo for acoustic event classification using a Convolutional Neural Network (CNN). When an acoustic event is fed as input into our system in real-time, the system performs the classification task and denotes to which class the acoustic event belongs. We combined different audio datasets into an own one consisting of 94 classes belonging to the context of Ambient Assisted Living (AAL). The so-called AAL-94 audio set is a combination of publicly available ESC-50 [7], Audio Set [4] and Ultrasound-8k [8] datasets. We enriched these subsets with own laboratory recordings to create a collection of 18,882 audio recordings typical for AAL. The datasets were trained and the classification task is performed using a CNN. The best model from the training process has been snapshot and is used for real-time audio processing in our demo. The latter visualizes the audio classification results in a real-time spectrogram and some statistical plots. Users either interacts creating noises themselves from the 94 available classes shown on an auxiliary screen of the demo, or trigger sounds from a MIDI keyboard to test the system performance live. Current and overall classification results are demonstrated on the main screen.

Intelligent IoT-based ambient assisted living systems (AALS) have been a major research focus area in recent times. Application of machine learning in areas of AALS such as fall detection has the potential to have huge social impact. There has been active research in the application of machine learning in fall detection, using data generated by various means such as wearable devices, environment sensors and vision based systems. The main challenge is to create a model that detects falls accurately, while keeping the design of the fall detection system minimal and non-intrusive. Wearable devices equipped with inertial motion unit (IMU) sensors and vital signs sensors are commonly used to enable analysis around performance of machine learning (ML) models. In this paper, we analyze the impact of using IMU sensor parameters in combination with vital signs parameters, on the performance of ML algorithms for fall detection. We present details on the data set we have generated for this purpose, and compare the performance of various ML algorithms on the collected dataset, with features from IMU sensors vis-a-vis those from IMU sensors in combination with vital signs sensors. We also apply machine learning algorithms on two public datasets, one with only IMU sensor parameter values and the second with only vital signs parameter values, and summarize their performance.

Wandering is a common behavior during all stages of Alzheimer’s Disease (AD), but it can be distressing to both people with dementia and their caregivers. Ambient Assisted Living (AAL) solutions have been increasingly relevant to promote ageing in place, targeting both autonomous and dependent older adults. Despite the potentials of AAL solutions, their pervasiveness can raise several security and privacy challenges. With mobile mockups, this work presents a use-case with SoTRAACE - Socio-Technical Risk-Adaptable Access Control model, applied to the prevention of getting lost in AD. SoTRAACE functionalities are illustrated as embedded in an alarm navigation system aimed at monitoring the person’s geolocation and alerting when a predefined safety perimeter is crossed. The model, applied to a healthcare scenario, assesses security and privacy risks at each moment of interaction and provides an innovative way for improved adaptable security visualization in AAL environments.

Daily human activity recognition using sensor data can be a fundamental task for many real-world applications, such as home monitoring and assisted living. One of the challenges in human activity recognition is to distinguish activities that have infrequent occurrence and less distinctive patterns. We propose a dissimilarity representation-based hierarchical classifier to perform two-phase learning. In the first phase, the classifier learns general features to recognise majority classes, and the second phase is to collect minority and subtle classes to identify fine difference between them. We compare our approach with a collection of state-of-the-art classification techniques on a real-world third-party dataset that is collected in a two-user home setting. Our results demonstrate that our hierarchical classifier approach outperforms the existing techniques in distinguishing users in performing the same type of activities. The key novelty of our approach is the exploration of dissimilarity representations and hierarchical classifiers, which allows us to highlight the difference between activities with subtle difference, and thus allows the identification of well-discriminating features. Postprint

Indoor location at room level plays a key role for providing useful services for Ambient Assisted Living (AAL) applications. Wi-Fi fingerprinting indoor location methods are extensively used due to the widespread availability of Wi-Fi infrastructures. A main drawback of Wi-Fi fingerprinting methods is the temporal cost involved in creating the radio maps. Crowdsourcing strategies have been presented as a way to minimize the cost of radio map creation. In this work, we present an extensive study of the issues involved when using crowdsourcing strategies for that purpose. Results provided by extensive experiments performed in a real scenario by three users during two weeks are presented. The main conclusions are: i) crowdsourcing data improves accuracy location in most studied cases; ii) accuracy of Wi-Fi fingerprinting methods decay along time; iii) device diversity is an important issue even when using the same device model.

In this work we analyze the security requirements and challenges of e-Health Internet of Things (IoT) applications and propose a complete architecture to address them. This architecture combines VICINITY IoT Framework security features together with General Data Protection Regulation (GDPR) compliant mechanisms in order to provide secure e-Health services to elders and middle-aged people. We also demonstrate how an Ambient Assisted Living (AAL) and an mHealth application were designed and implemented, addressing the current security and privacy requirements.

Elderly people is increasing dramatically during the current years, and it is expected that this population reaches 2.1 billion of individuals by 2050. In this regard, new care strategies are required. Assisted living technologies have proposed alternatives to support professional caregivers and families to take care of elderly people, such as in risk of falls. Currently, fall detection systems are able to alleviate the latter problem and reduce the time a person who suffered a fall receives assistance. Thus, this paper proposes a fall detection system based on image processing strategy to extract motion features through an optical flow method. For classification, we use these features as inputs to a convolutional neural network. We applied our approach in a dataset comprises video recordings of one subject performing different types of falls. In experimental results, our approach showed 92% accuracy on the dataset used.

The considerable rise in the average age of the population leads to the exceeding number of seniority people. This has made a strong impact in demanding community requirement such as rehabilitation, physical support, home assistance etc. For this reason, Ambient Assisted Living (AAL) plays an exceptional role in supervising the Activities of Daily Living (ADL) of the target users. AAL would be an encouraging technology for the current care models by acting as a companion. This becomes a thought-provoking research area in the fast-developing world, but the condition for examining various ADL and the self-classification becomes a big challenge. There are numerous ways to accumulate data from camera, microphone and other sensors. Datasets that are used in this paper have been collected from publicly available resources for Human Activity Recognition. In this work, we have discussed about different classification algorithm used for grouping of features into their corresponding activity and by then their performance is evaluated. However, many classifiers face the limitations of enormous number of datasets and immense training time of the feature vector. Hence to overcome these problems, we have used RF and K-NN classifiers. Both are tremendously helpful in dealing with larger datasets and provides an accuracy level of about 83.05%.

With the increasing number of scientific publications, the analysis of the trends and the state-of-the-art in a certain scientific field is becoming very time-consuming and tedious task. In response to urgent needs of information, for which the existing systematic review model does not well, several other review types have emerged, namely the rapid review and scoping reviews. In this paper, we propose an NLP powered tool that automates most of the review process by automatic analysis of articles indexed in the IEEE Xplore, PubMed, and Springer digital libraries. We demonstrate the applicability of the toolkit by analyzing articles related to Enhanced Living Environments and Ambient Assisted Living, in accordance with the PRISMA surveying methodology. The relevant articles were processed by the NLP toolkit to identify articles that contain up to 20 properties clustered into 4 logical groups. The analysis showed increasing attention from the scientific communities towards Enhanced and Assisted living environments over the last 10 years and showed several trends in the specific research topics that fall into this scope. The case study demonstrates that the NLP toolkit can ease and speed up the review process and show valuable insights from the surveyed articles even without manually reading of most of the articles. Moreover, it pinpoints the most relevant articles which contain more properties and therefore, significantly reduces the manual work, while also generating informative tables, charts and graphs.

This chapter introduces an ontology-based knowledge-driven approach to real-time, continuous activity recognition based on multi-sensor data streams in the context of assisted living within smart homes. It first presents a generic system architecture for the proposed knowledge-driven approach and its underlying ontology-based activity recognition process. It then analyses the characteristics of smart homes and Activities of Daily Living (ADL) upon which both context and ADL ontologies are developed. Following this, the chapter describes algorithms for activity recognition based on semantic subsumption reasoning. Finally, an example case study is conducted using an implemented function-rich software system, which evaluates and demonstrates the proposed approach through extensive experiments involving a number of various ADL use scenarios.

While the increasing average age of population is posing new challenges to societies and healthcare systems, the emergence of the Internet of Things research area is generating the hope for automated assisted environments, which could combine the advances in sensors networks with that of runtime monitoring systems, in order to create smart houses able to take care of their older inhabitants and delay the recourse to hospitals and nursing homes. However, although various assisted living systems have been proposed in the last decade, the goal of realizing an effective domestic support system for elderly is still far from reached. In this work, we present a project aiming to re-engineer a set of everyday life objects, equipping them with environmental and wearable sensors, thus to monitor the condition of older people in their domestic residences and provide security while preserving the autonomy and independence of the subjects. The main focus of the paper at hand is on the requirements and solutions implemented to realize the backbone infrastructure of such system as regards both the adopted semantic message routing mechanism and the newly conceived approach to event analysis, which combines Complex Event Processing and a reactive implementation of Event Calculus.

In the time of an ageing world, aging at home is both an economically and socially viable and sustainable solution which may also improve the elderly’s well-being. There are currently a few Home-based Elderly Care systems which although operational are not yet optimal in terms of efficiency and security. The paper propose a Home-based Elderly Care solution which makes use of the 5G mobile network slicing with two secure and isolated slices, namely the Health Care slice and the Smart Home slice to provide an inherent secure connection. Further, the solution includes an efficient and secure Emergency Call which ensures that the appropriate caregivers can dispatched and provide help in shorter times. A proof-of-concept implementation is described thoroughly. This paper is a result of the SCOTT project (www.scott-project.eu) which has received funding from the Electronic Component Systems for European Leadership Joint Undertaking under grant agreement No. 737422. This Joint Undertaking receives support from the EU H2020 research and innovation programme and Austria, Spain, Finland, Ireland, Sweden, Germany, Poland, Portugal, Netherlands, Belgium, Norway.