Design and Evaluation of Personalized Services to Foster Active Aging: The Experience of Technology Pre-Validation in Italian Pilots.

Assistive devices could promote independent living and support the active and healthy aging of an older population; however, several factors can badly influence the long-term use of new technologies. In this context, this paper presents a two-step methodology called "pre-validation" that aims to identify the factors that can bias the use of new services, thus minimizing the risk of an unsuccessful longer trial. The proposed pre-validation methodology is composed of two main phases that aim to assess the usability and the reliability of the technology assessed in a laboratory environment and the usability, acceptability, user experience, and reliability of the technology in real environments. The tested services include the socialization scenario, in which older adults are better connected to the community via technological solutions (i.e., socialization applications), and the monitoring scenario, which allows for the introduction of timely interventions (technologies involved include environmental monitoring sensors, a telepresence robot, wearable sensors, and a personalized dashboard). The obtained results underline an acceptable usability level (average System Usability Scale score > 65) for the tested technologies (i.e., socialization applications and a telepresence robot). Phase Two also underlines the good acceptability, user experience, and usability of the tested services. The statistical analysis underlines a correlation between the stress related to the use of technology, digital skills, and intention of use, among other factors. Qualitative feedback also remarks on a correlation between older adults with low digital skills and an anxiety about using technology. Positive correlation indexes were highlighted between the trust and usability scores. Eventually, future long-term trials with assistive technology should rely on motivated caregivers, be founded on a strong recruitment process, and should reassure older adults­especially the ones with low digital literacy­about the use of technology by proposing personalized training and mentoring, if necessary, to increase the trust.

Hands-feet wireless devices: Test-retest reliability and discriminant validity of motor measures in Parkinson's disease telemonitoring.

BACKGROUND: Telemonitoring, a branch of telemedicine, involves the use of technological tools to remotely detect clinical data and evaluate patients. Telemonitoring of patients with Parkinson's disease (PD) should be performed using reliable and discriminant motor measures. Furthermore, the method of data collection and transmission, and the type of subjects suitable for telemonitoring must be well defined. OBJECTIVE: To analyze differences in patients with PD and healthy controls (HC) with the wearable inertial device SensHands-SensFeet (SH-SF), adopting a standardized acquisition mode, to verify if motor measures provided by SH-SF have a high discriminating capacity and high intraclass correlation coefficient (ICC). METHODS: Altogether, 64 patients with mild-to-moderate PD and 50 HC performed 14 standardized motor activities for assessing bradykinesia, postural and resting tremors, and gait parameters. SH-SF inertial devices were used to acquire movements and calculate objective motor measures of movement (total: 75). For each motor task, five or more biomechanical parameters were measured twice. The results were compared between patients with PD and HC. RESULTS: Fifty-eight objective motor measures significantly differed between patients with PD and HC; among these, 32 demonstrated relevant discrimination power (Cohen's d > 0.8). The test-retest reliability was excellent in patients with PD (median ICC = 0.85 right limbs, 0.91 left limbs) and HC (median ICC = 0.78 right limbs, 0.82 left limbs). CONCLUSION: In a supervised environment, the SH-SF device provides motor measures with good results in terms of reliability and discriminant ability. The reliability of SH-SF measurements should be evaluated in an unsupervised home setting in future studies.

Learning-based control approaches for service robots on cloth manipulation and dressing assistance: a comprehensive review.

BACKGROUND: Service robots are defined as reprogrammable, sensor-based mechatronic devices that perform useful services in an autonomous or semi-autonomous way to human activities in an everyday environment. As the number of elderly people grows, service robots, which can operate complex tasks like dressing tasks for disabled people, are being demanded increasingly. Consequently, there is a growing interest in studying dressing tasks, such as putting on a t-shirt, a hat, or shoes. Service robots or robot manipulators have been developed to accomplish these tasks using several control approaches. The robots used in this kind of application are usually bimanual manipulator (i.e. Baxter robot) or single manipulators (i.e. Ur5 robot). These arms are usually used for recognizing clothes and then folding them or putting an item on the arm or on the head of a person. METHODS: This work provides a comprehensive review of the most relevant attempts/works of robotic dressing assistance with a focus on the control methodology used for dressing tasks. Three main areas of control methods for dressing tasks are proposed: Supervised Learning (SL), Learning from Demonstration (LfD), and Reinforcement Learning (RL). There are also other methods that cannot be classified into these three areas and hence they have been placed in the other methods section. This research was conducted within three databases: Scopus, Web of Science, and Google Scholar. Accurate exclusion criteria were applied to screen the 2594 articles found (at the end 39 articles were selected). For each work, an evaluation of the model is made. CONCLUSION: Current research in cloth manipulation and dressing assistance focuses on learning-based robot control approach. Inferring the cloth state is integral to learning the manipulation and current research uses principles of Computer Vision to address the issue. This makes the larger problem of control robot based on learning data-intensive; therefore, a pressing need for standardized datasets representing different cloth shapes, types, materials, and human demonstrations (for LfD) exists. Simultaneously, efficient simulation capabilities, which closely model the deformation of clothes, are required to bridge the reality gap between the real-world and virtual environments for deploying the RL trial and error paradigm. Such powerful simulators are also vital to collect valuable data to train SL and LfD algorithms that will help reduce human workload.

Image Classification Using Multiple Convolutional Neural Networks on the Fashion-MNIST Dataset.

As the elderly population grows, there is a need for caregivers, which may become unsustainable for society. In this situation, the demand for automated help increases. One of the solutions is service robotics, in which robots have automation and show significant promise in working with people. In particular, household settings and aged people's homes will need these robots to perform daily activities. Clothing manipulation is a daily activity and represents a challenging area for a robot. The detection and classification are key points for the manipulation of clothes. For this reason, in this paper, we proposed to study fashion image classification with four different neural network models to improve apparel image classification accuracy on the Fashion-MNIST dataset. The network models are tested with the highest accuracy with a Fashion-Product dataset and a customized dataset. The results show that one of our models, the Multiple Convolutional Neural Network including 15 convolutional layers (MCNN15), boosted the state of art accuracy, and it obtained a classification accuracy of 94.04% on the Fashion-MNIST dataset with respect to the literature. Moreover, MCNN15, with the Fashion-Product dataset and the household dataset, obtained 60% and 40% accuracy, respectively.

Emotion Recognizing by a Robotic Solution Initiative (EMOTIVE Project).

BACKGROUND: Emotion recognition skills are predicted to be fundamental features in social robots. Since facial detection and recognition algorithms are compute-intensive operations, it needs to identify methods that can parallelize the algorithmic operations for large-scale information exchange in real time. The study aims were to identify if traditional machine learning algorithms could be used to assess every user emotions separately, to relate emotion recognizing in two robotic modalities: static or motion robot, and to evaluate the acceptability and usability of assistive robot from an end-user point of view. METHODS: Twenty-seven hospital employees (M = 12; F = 15) were recruited to perform the experiment showing 60 positive, negative, or neutral images selected in the International Affective Picture System (IAPS) database. The experiment was performed with the Pepper robot. Concerning experimental phase with Pepper in active mode, a concordant mimicry was programmed based on types of images (positive, negative, and neutral). During the experimentation, the images were shown by a tablet on robot chest and a web interface lasting 7 s for each slide. For each image, the participants were asked to perform a subjective assessment of the perceived emotional experience using the Self-Assessment Manikin (SAM). After participants used robotic solution, Almere model questionnaire (AMQ) and system usability scale (SUS) were administered to assess acceptability, usability, and functionality of robotic solution. Analysis wasperformed on video recordings. The evaluation of three types of attitude (positive, negative, andneutral) wasperformed through two classification algorithms of machine learning: k-nearest neighbors (KNN) and random forest (RF). RESULTS: According to the analysis of emotions performed on the recorded videos, RF algorithm performance wasbetter in terms of accuracy (mean ± sd = 0.98 ± 0.01) and execution time (mean ± sd = 5.73 ± 0.86 s) than KNN algorithm. By RF algorithm, all neutral, positive and negative attitudes had an equal and high precision (mean = 0.98) and F-measure (mean = 0.98). Most of the participants confirmed a high level of usability and acceptability of the robotic solution. CONCLUSIONS: RF algorithm performance was better in terms of accuracy and execution time than KNN algorithm. The robot was not a disturbing factor in the arousal of emotions.

On the Use of Assistive Technology during the COVID-19 Outbreak: Results and Lessons Learned from Pilot Studies.

As a consequence of the COVID-19 emergency, frail citizens felt isolated because of social isolation, suspended and/or strongly reduced home assistance, and limited access to hospitals. In this sense, assistive technology could play a pivotal role in empowering frail older adults reducing their isolation, as well as in reinforcing the work of formal caregivers and professionals. In this context, the goal of this paper is to present four pilot studies-conducted from March 2020 to April 2021-to promptly react to COVID-19 by providing assistive technology solutions, aiming to (1) guarantee high-quality service to older adults in-home or in residential facility contexts, (2) promote social inclusion, and (3) reduce the virus transmission. In particular, four services, namely, telepresence service, remote monitoring service, virtual visit, and environmental disinfection, were designed, implemented, and tested in real environments involving 85 end-users to assess the user experience and/or preliminary assess the technical feasibility. The results underlined that all the proposed services were generally accepted by older adults and professionals. Additionally, the results remarked that the use of telepresence robots in private homes and residential facilities increased enjoyment reducing anxiety, whereas the monitoring service supported the clinicians in monitoring the discharged COVID-19 patients. It is also worth mentioning that two new services/products were developed to disinfect the environment and to allow virtual visits within the framework of a hospital information system. The virtual visits service offered the opportunity to expand the portfolio of hospital services. The main barriers were found in education, technology interoperability, and ethical/legal/privacy compliance. It is also worth mentioning the key role played by an appropriate design and customer needs analysis since not all assistive devices were designed for older persons.

A wearable ring-shaped inertial system to identify action planning impairments during reach-to-grasp sequences: a pilot study.

BACKGROUND: The progressive ageing of the population is leading to an increasing number of people affected by cognitive decline, including disorders in executive functions (EFs), such as action planning. Current procedures to evaluate cognitive decline are based on neuropsychological tests, but novel methods and approaches start to be investigated. Reach-to-grasp (RG) protocols have shown that intentions can influence the EFs of action planning. In this work, we proposed a novel ring-shaped wearable inertial device, SensRing, to measure kinematic parameters during RG and after-grasp (AG) tasks with different end-goals. The aim is to evaluate whether SensRing can characterize the motor performances of people affected by Mild Neurocognitive Disorder (MND) with impairment in EFs. METHODS: Eight Individuals with dysexecutive MND, named d-MND, were compared to ten older healthy subjects (HC). They were asked to reach and grasp a can with three different intentions: to drink (DRINK), to place it on a target (PLACE), or to pass it to a partner (PASS). Twenty-one kinematic parameters were extracted from SensRing inertial data. RESULTS: Seven parameters resulted able to differentiate between HC and d-MND in the RG phase, and 8 features resulted significant in the AG phase. d-MND, indeed, had longer reaction times (in RG PLACE), slower peak velocities (in RG PLACE and PASS, in AG DRINK and PLACE), longer deceleration phases (in all RG and AG DRINK), and higher variability (in RG PLACE, in AG DRINK and PASS). Furthermore, d-MND showed no significant differences among conditions, suggesting that impairments in EFs influence their capabilities in modulating the action planning based on the end-goal. CONCLUSIONS: Based on this explorative study, the system might have the potential for objectifying the clinical assessment of people affected by d-MND by administering an easy motor test. Although these preliminary results have to be investigated in-depth in a larger sample, the portability, wearability, accuracy, and ease-of use of the system make the SensRing potentially appliable for remote applications at home, including analysis of protocols for neuromotor rehabilitation in patients affected by MND.

Pilots for Healthy and Active Ageing (PHArA-ON) Project: Definition of New Technological Solutions for Older People in Italian Pilot Sites Based on Elicited User Needs.

BACKGROUND: The Pilots for Healthy and Active Ageing (PHArA-ON) project aimsto ensure reality smart and active living for Europe's ageing population by creating a set of integrated and highly customizable interoperable open platforms with advanced services, devices, and technologies and tools. The aim of the present study was to determine the needs and preferences of older people and their caregivers for improving healthy and active aging and guiding the technological development of thePHArA-ON system. METHODS: A pre-structured interview was administered to older adults, informal caregivers and professional caregivers (including social operators) taking part in the piloting sessions. RESULTS: Interviews were carried out in Umana Persone Social Enterprise R&D Network (UP) in Tuscany, and Ospedale Casa SollievodellaSofferenza (CSS) in Apulia. A total of 22 older adults, 22 informal caregivers, 13 professional caregivers and 4 social operators were recruited. A prioritization analysis of services, according to the stakeholder's needs, has determined two fundamental need categories: Heath Management (i.e., stimulation and monitoring), and Socialisation (i.e., promoting social inclusion). CONCLUSIONS: The main scientific contributions to this study are the following: to design and evaluate technology in the context of healthy and active ageing, to acquire relevant knowledge on user needs to develop technologies that can handle the real life situations of older people, obtain useful insights about the attitude and availability of end-users in using technologies in clinical practice, and to provide important guidelines to improve the PHArA-ON system. Specific experimentation stages were also carried out to understand which kind of technology is more acceptable, and to obtain feedback regarding the development priority related to the impact of the proposed services. Research through fruitful and continuous interaction with the different subjects involved in the development process of the system, as well as with stakeholders, enabled the implementation of a platform which could be further and easily integrated and improved.

A Wearable System to Objectify Assessment of Motor Tasks for Supporting Parkinson's Disease Diagnosis.

Objective assessment of the motor evaluation test for Parkinson's disease (PD) diagnosis is an open issue both for clinical and technical experts since it could improve current clinical practice with benefits both for patients and healthcare systems. In this work, a wearable system composed of four inertial devices (two SensHand and two SensFoot), and related processing algorithms for extracting parameters from limbs motion was tested on 40 healthy subjects and 40 PD patients. Seventy-eight and 96 kinematic parameters were measured from lower and upper limbs, respectively. Statistical and correlation analysis allowed to define four datasets that were used to train and test five supervised learning classifiers. Excellent discrimination between the two groups was obtained with all the classifiers (average accuracy ranging from 0.936 to 0.960) and all the datasets (average accuracy ranging from 0.953 to 0.966), over three conditions that included parameters derived from lower, upper or all limbs. The best performances (accuracy = 1.00) were obtained when classifying all the limbs with linear support vector machine (SVM) or gaussian SVM. Even if further studies should be done, the current results are strongly promising to improve this system as a support tool for clinicians in objectifying PD diagnosis and monitoring.

The Rehapiano-Detecting, Measuring, and Analyzing Action Tremor Using Strain Gauges.

We have developed a device, the Rehapiano, for the fast and quantitative assessment of action tremor. It uses strain gauges to measure force exerted by individual fingers. This article verifies the device's capability to measure and monitor the development of upper limb tremor. The Rehapiano uses a precision, 24-bit, analog-to-digital converter and an Arduino microcomputer to transfer raw data via a USB interface to a computer for processing, database storage, and evaluation. First, our experiments validated the device by measuring simulated tremors with known frequencies. Second, we created a measurement protocol, which we used to measure and compare healthy patients and patients with Parkinson's disease. Finally, we evaluated the repeatability of a quantitative assessment. We verified our hypothesis that the Rehapiano is able to detect force changes, and our experimental results confirmed that our system is capable of measuring action tremor. The Rehapiano is also sensitive enough to enable the quantification of Parkinsonian tremors.

Assistive robots for socialization in elderly people: results pertaining to the needs of the users.

BACKGROUND/AIM: Technological solutions can support the elderly, improve their quality of life and reduce isolation and loneliness. The Euro-Japan ACCRA (Agile Co-Creation for Robots and Aging) project has the objective of building a reference co-creation methodology for the development of robotic solutions for ageing. The aim of this study is to provide a pilot qualitative analysis of the real needs of elderly people and their caregivers when exposed to conversational activities with robots and to identify priority needs that should be developed from end-user perspectives. METHODS: A qualitative research design was adopted to define a pre-structured questionnaire that was administered to the elderly taking part in the piloting sessions. Three groups of end-users were included: subjects with an age ≥ 60 years, informal caregivers and formal caregivers. RESULTS: The interviews were carried out in Italy and Japan. A total of 17 elderly and 36 caregivers were recruited. Common needs in the two sites were categorized into 3 groups: Communication; Emotion Detection and Safety. General robot acceptance level is good and perception is positive among participants in the pilot sites. CONCLUSION: A positive perception of the elderly on the application of a robotic solution was found and many are the needs that could be addressed by an appropriate and careful robotic development taking into account the real needs and capabilities of the involved subjects.

Can physical and cognitive training based on episodic memory be combined in a new protocol for daily training?

BACKGROUND: Cognitive training (CT) is defined as guided practice on a set of standard tasks designed to stimulate particular cognitive functions. Recent studies have shown that physical exercise is beneficial for cognitive activity in older adults and patients with degenerative diseases. AIMS: The main objective of the present study is to create a new cognitive tool able to provide training for cognitive functions that take advantage of the physical activity involved in the execution of the task. A study concerning the application of a new CT tool for episodic memory is presented and divided in two parts. The first one aims at developing a new sensorized device, called SmartTapestry, for physical and cognitive training. The second part aims at understanding its technical viability and level of sensitivity in stimulating the same cognitive domain covered by the standardized tests, despite the introduction of the physical activity variable. METHODS: The SmartTapestry device was tested with a total of 53 subjects, 29 healthy subjects and 24 subjects suffering from mild cognitive impairment. RESULTS AND DISCUSSIONS: The results show a good correlation between the two approaches (p < 0.005), suggesting that SmartTapestry can stimulate the same cognitive functions of traditional cognitive tasks, with the addition of physical exercise. CONCLUSIONS: The results of this study may be useful in designing ecological and combined cognitive-physical tools, which can be used daily at home, reducing the presence of clinical staff, to train at the same time the brain and the body so as to improve the cognitive treatments efficacy.

Correction: Fiorini, L., et al., Unsupervised Machine Learning for Developing Personalised Behaviour Models Using Activity Data. Sensors 2017, 17, 1034.

A correction is presented to correct the section headings of Sections 5.1, 5.2, and 5.3 in[Sensors, 2017, 17, 1034].

Psychophysiological Indicators for Modeling User Experience in Interactive Digital Entertainment.

Analyses of user experience in the electronic entertainment industry currently rely on self-reporting methods, such as surveys, ratings, focus group interviews, etc. We argue that self-reporting alone carries inherent problems-mainly the misinterpretation and temporal delay during longer experiments-and therefore, should not be used as a sole metric. To tackle this problem, we propose the possibility of modeling consumer experience using psychophysiological measures and demonstrate how such models can be trained using machine learning methods. We use a machine learning approach to model user experience using real-time data produced by the autonomic nervous system and involuntary psychophysiological responses. Multiple psychophysiological measures, such as heart rate, electrodermal activity, and respiratory activity, have been used in combination with self-reporting to prepare training sets for machine learning algorithms. The training data was collected from 31 participants during hour-long experiment sessions, where they played multiple video-games. Afterwards, we trained and compared the results of four different machine learning models, out of which the best one produced ∼96% accuracy. The results suggest that psychophysiological measures can indeed be used to assess the enjoyment of digital entertainment consumers.

Automated Systems Based on Wearable Sensors for the Management of Parkinson's Disease at Home: A Systematic Review.

BACKGROUND: Parkinson's disease is a common neurodegenerative pathology that significantly influences quality of life (QoL) of people affected. The increasing interest and development in telemedicine services and internet of things technologies aim to implement automated smart systems for remote assistance of patients. The wide variability of Parkinson's disease in the clinical expression, as well as in the symptom progression, seems to address the patients' care toward a personalized therapy. OBJECTIVES: This review addresses automated systems based on wearable/portable devices for the remote treatment and management of Parkinson's disease. The idea is to obtain an overview of the telehealth and automated systems currently developed to address the impairments due to the pathology to allow clinicians to improve the quality of care for Parkinson's disease with benefits for patients in QoL. DATA SOURCES: The research was conducted within three databases: IEEE Xplore®, Web of Science®, and PubMed Central®, between January 2008 and September 2017. STUDY ELIGIBILITY CRITERIA: Accurate exclusion criteria and selection strategy were applied to screen the 173 articles found. RESULTS: Ultimately, 55 articles were fully evaluated and included in this review. Divided into three categories, they were automated systems actually tested at home, implemented mobile applications for Parkinson's disease assessment, or described a telehealth system architecture. CONCLUSION: This review would provide an exhaustive overview of wearable systems for the remote management and automated assessment of Parkinson's disease, taking into account the reliability and acceptability of the implemented technologies.

Evaluation of a Companion Robot for Individuals With Dementia: Quantitative Findings of the MARIO Project in an Irish Residential Care Setting.

The current study focuses on the short-term effect of MARIO, a social robot, on quality of life, depression, and perceived social support in persons with dementia (PWD) and evaluates their acceptability of MARIO. Ten PWD in one nursing home took part in a 4-week pilot study, where each participant had up to 12 sessions with MARIO. Sessions comprised engagement in music, news, reminiscence, games, and calendar applications. Standardized questionnaires were administered before and after the 4-week period. Participants had a sustained interest in MARIO during their interactions and an acceptance of MARIO's appearance, sound, and applications. Consequently, participants spent more time socially engaged. No statistically significant differences were found in quality of life, depression, and perceived social support. PWD can engage with a social robot in a real-world nursing home. Future research should incorporate a larger sample and longer intervention period. [Journal of Gerontological Nursing, 45(7), 36-45.].

Robotic Services Acceptance in Smart Environments With Older Adults: User Satisfaction and Acceptability Study.

BACKGROUND: In Europe, the population of older people is increasing rapidly. Many older people prefer to remain in their homes but living alone could be a risk for their safety. In this context, robotics and other emerging technologies are increasingly proposed as potential solutions to this societal concern. However, one-third of all assistive technologies are abandoned within one year of use because the end users do not accept them. OBJECTIVE: The aim of this study is to investigate the acceptance of the Robot-Era system, which provides robotic services to permit older people to remain in their homes. METHODS: Six robotic services were tested by 35 older users. The experiments were conducted in three different environments: private home, condominium, and outdoor sites. The appearance questionnaire was developed to collect the users' first impressions about the Robot-Era system, whereas the acceptance was evaluated through a questionnaire developed ad hoc for Robot-Era. RESULTS: A total of 45 older users were recruited. The people were grouped in two samples of 35 participants, according to their availability. Participants had a positive impression of Robot-Era robots, as reflected by the mean score of 73.04 (SD 11.80) for DORO's (domestic robot) appearance, 76.85 (SD 12.01) for CORO (condominium robot), and 75.93 (SD 11.67) for ORO (outdoor robot). Men gave ORO's appearance an overall score higher than women (P=.02). Moreover, participants younger than 75 years understood more readily the functionalities of Robot-Era robots compared to older people (P=.007 for DORO, P=.001 for CORO, and P=.046 for ORO). For the ad hoc questionnaire, the mean overall score was higher than 80 out of 100 points for all Robot-Era services. Older persons with a high educational level gave Robot-Era services a higher score than those with a low level of education (shopping: P=.04; garbage: P=.047; reminding: P=.04; indoor walking support: P=.006; outdoor walking support: P=.03). A higher score was given by male older adults for shopping (P=.02), indoor walking support (P=.02), and outdoor walking support (P=.03). CONCLUSIONS: Based on the feedback given by the end users, the Robot-Era system has the potential to be developed as a socially acceptable and believable provider of robotic services to facilitate older people to live independently in their homes.

Smartphone-Based Food Diagnostic Technologies: A Review.

A new generation of mobile sensing approaches offers significant advantages over traditional platforms in terms of test speed, control, low cost, ease-of-operation, and data management, and requires minimal equipment and user involvement. The marriage of novel sensing technologies with cellphones enables the development of powerful lab-on-smartphone platforms for many important applications including medical diagnosis, environmental monitoring, and food safety analysis. This paper reviews the recent advancements and developments in the field of smartphone-based food diagnostic technologies, with an emphasis on custom modules to enhance smartphone sensing capabilities. These devices typically comprise multiple components such as detectors, sample processors, disposable chips, batteries and software, which are integrated with a commercial smartphone. One of the most important aspects of developing these systems is the integration of these components onto a compact and lightweight platform that requires minimal power. To date, researchers have demonstrated several promising approaches employing various sensing techniques and device configurations. We aim to provide a systematic classification according to the detection strategy, providing a critical discussion of strengths and weaknesses. We have also extended the analysis to the food scanning devices that are increasingly populating the Internet of Things (IoT) market, demonstrating how this field is indeed promising, as the research outputs are quickly capitalized on new start-up companies.

A Human Activity Recognition System Based on Dynamic Clustering of Skeleton Data.

Human activity recognition is an important area in computer vision, with its wide range of applications including ambient assisted living. In this paper, an activity recognition system based on skeleton data extracted from a depth camera is presented. The system makes use of machine learning techniques to classify the actions that are described with a set of a few basic postures. The training phase creates several models related to the number of clustered postures by means of a multiclass Support Vector Machine (SVM), trained with Sequential Minimal Optimization (SMO). The classification phase adopts the X-means algorithm to find the optimal number of clusters dynamically. The contribution of the paper is twofold. The first aim is to perform activity recognition employing features based on a small number of informative postures, extracted independently from each activity instance; secondly, it aims to assess the minimum number of frames needed for an adequate classification. The system is evaluated on two publicly available datasets, the Cornell Activity Dataset (CAD-60) and the Telecommunication Systems Team (TST) Fall detection dataset. The number of clusters needed to model each instance ranges from two to four elements. The proposed approach reaches excellent performances using only about 4 s of input data (~100 frames) and outperforms the state of the art when it uses approximately 500 frames on the CAD-60 dataset. The results are promising for the test in real context.

Unsupervised Machine Learning for Developing Personalised Behaviour Models Using Activity Data.

The goal of this study is to address two major issues that undermine the large scale deployment of smart home sensing solutions in people's homes. These include the costs associated with having to install and maintain a large number of sensors, and the pragmatics of annotating numerous sensor data streams for activity classification. Our aim was therefore to propose a method to describe individual users' behavioural patterns starting from unannotated data analysis of a minimal number of sensors and a "blind" approach for activity recognition. The methodology included processing and analysing sensor data from 17 older adults living in community-based housing to extract activity information at different times of the day. The findings illustrate that 55 days of sensor data from a sensor configuration comprising three sensors, and extracting appropriate features including a "busyness" measure, are adequate to build robust models which can be used for clustering individuals based on their behaviour patterns with a high degree of accuracy (>85%). The obtained clusters can be used to describe individual behaviour over different times of the day. This approach suggests a scalable solution to support optimising the personalisation of care by utilising low-cost sensing and analysis. This approach could be used to track a person's needs over time and fine-tune their care plan on an ongoing basis in a cost-effective manner.