Chemical polymerization of o-phenylenediamine (OPD) in the presence of poly(ethylene oxide), double-wall carbon nanotubes (DWNTs) and ferric chloride is carried out in order to obtain composites based on the poly(o-phenylenediamine)-poly(ethylene oxide) (POPD-PEO) fibres covered and interconnected with DWNTs. Vibrational and photoluminescence properties of these composite materials as well as their morphologies are shown by infrared (IR) spectroscopy, Raman scattering, photoluminescence (PL) and scanning electron microscopy (SEM). An adsorption of DWNTs onto the POPD rods surface in the absence and in the presence of PEO is highlighted by SEM. The vibrational changes reported by Raman scattering and IR spectroscopy prove a covalent functionalization of DWNTs with the macromolecular compound POPD which is doped with FeCl-₄ ions. New hydrogen bonds are generated between POPD covalently functionalized DWNTs and hydroxyl groups of PEO according to IR spectroscopic studies. The two macromolecular compounds, POPD and POPD-PEO, show a complex emission band with maxima at 572 and 566 nm, having a shoulder at 667 nm. A significant change in the profile of the PL bands of POPD and POPD-PEO is induced in the DWNTs presence. We show that DWNTs induce (i) a diminution in the POPD PL band intensity peaked between 525-600 nm simultaneous with the increase in the intensity of the PL band situated in the 600-800 nm spectral range and (ii) an enhancement process of the emission band localized in the 475-800 nm spectral range in the case of POPD-PEO.

The goal of the proposed system is to retrieve the corresponding image from the database based on the query image. Now-a-days images are stored in the database in the form of digital. Thus, retrieval of image from huge database become complex. Most of the existing system uses indirect method of retrieval and they have no methodology. Thus the major aim of our approach is to construct an effective and efficient search engine tool in order to retrieve image from a huge database based on the user query.

Sleep represents a main factor for a healthy lifestyle. It improves health and quality of life in many ways: 1) it is vital for physical health; 2) it is essential for mental health; 3) improves memory and focus and 4) promotes personal and public safety. Sleep is a basic biological function that undergoes changes along with ageing and many pathological conditions. Some changes due to age are so deep that it is hard to distinguish between ageing and disease. Each person is different, considering the activities they perform that influence the sleep during the night, but also their metabolism, age etc. Different methods are therefore needed to explain the variation in sleep patterns for many people. Also, physical activity and sleep are highly interrelated health behaviours. The physical activity during the day influences the quality of sleep, and vice versa. The purpose of this paper is to present a system for creating a user profile based on his sleep behaviour. Based on the learnt profile, the system will automatically detect changes in the user's sleep pattern. If there are changes, the system will automatically check other medical parameters: heart rate, respiratory rate, physical activity during day (eg. number of steps, number of stairs). In case of existing abnormal values, an alarm message will be sent to the user with some suggestions regarding his daily programme. Data analysed is collected using two sensors: Emfit QS sensor for sleep monitoring and Fitbit for physical activity monitoring. In case of sleep, different machine learning methods were applied through pairs of sleep parameters (eg duration in rem versus total sleep duration, duration in light versus total sleep duration, duration awake versus total sleep duration, activity versus toss and turns count) in order to learn the profile of the user. The learning method is based on regression models. The regression is performed on a variable number of days in order to find a model that will learn the data set. For each number of days used for training, the model score will be calculated on the following days, and in the end, the model that has managed to obtain the highest score will be chosen. The models used are simple linear regression, polynomial regression, regression trees and regression support vectors. The models were evaluated on data collected from three users. The data sets used for the evaluation of each user was collected over a period of six months. The following models have proven to be common to all users: the model that learns the correlation between the variation of sleep duration in the light stage versus the total sleep duration. Thus, the proposed system will be able to generate alerts or recommendations using his sleep pattern. In case of differences between his history, the user will receive information that will help him to correct his lifestyle. Through the learning profile of the user, the system will proactive supervise him, making possible to reduce the possibility of illness.

Human activity recognition has been a branch of interest in the field of computer vision for decades, due to its numerous applications in different domains, such as medicine, surveillance, entertainment or human-computer interaction. We propose an intuitive, effective, quickly trainable and customizable system for recognizing human activities designed with an automated machine learning method based on Neural Architecture Search. Information from all channels of a 3D video (RGB and depth data, skeleton and context objects) is merged by independently passing these data streams through 2D convolutional neural networks. The outputs of all networks are combined in a summarizing array of class scores using fusion mechanisms that are not computationally intensive but reflect the meaningful information from a video. The proposed system is tested using three public datasets and a new dataset-PRECIS HAR-that was created in our laboratory. In all our experiments, the system is proven to be highly accurate: 98.43% on MSRDailyActivity3D, 91.41% on UTD-MHAD, 90.95% on NTU RGB+D, and 94.38% on our dataset.

This paper proposes a set of serious games as a learning platform for cognitive training with the help of virtual reality. We plan to see how efficient is a serious game with virtual reality as a learning platform in this direction. The games have simple functioning mechanics. For example, one type of game is the following: it introduces the user into a natural scene using virtual reality. In the scene there are different objects with different sizes, colors and orientations and the objects will move through the scene. The user must collect objects in order to create a story with them or to create a picture composed of the collected objects (similar with a puzzle). To make the game more attractive for each user, the designing of the game is adapted based on both the user profile and the user performance. The scene will be adapted to the user profile and his preferences - each user will have a profile created when game starts. For example, if the user likes to climb the mountains, he will be placed into a mountain based scene. Also, the performance of the user will influence the designing and the difficulty of the game level: if the player is successful, the difficulty level will rise, otherwise the difficulty level will be decreased. Machine learning techniques will be considered for developing behaviour of the characters - especially for the space exploration, e.g. reinforcement learning techniques. We expect that the proposed games can also help teachers to: i) implement methodological improvements, ii) prevent of poor school performance and also iii) to correct learning problems.

The presented prototype for image acquisitions and processing (IPA) aims to create a maintenance road system with minimal cost, mounted on non-specialized vehicle, enabling image acquisition in various conditions. IPA has an important role in the proposed platform designed by PAV3M for intelligent management, monitoring and maintenance of pavements and roads. We have developed new image processing solutions, analysis methods and enhanced (more robust, efficient, dedicated) solutions for solving the specific problems related to pavement analysis using PCI standards for road crack detection and classification.

Social assistive robotics is at the forefront of the effort for ensuring independent living and inclusion, especially for the elderly and vulnerable members of the population. We consider that to be proactive is a core requirement of such robotic systems. One key aspect of being proactive is to understand what the user is doing. We explored the task of human action recognition in the context of a social robotic platform based on a framework that we proposed in earlier work. We implemented a processing pipeline specific to this task then collected a dataset using a Pepper robot. The proposed domain adaptation method to increase performance of the method showed promising results.

Recent developments in robotics and machine learning make robot assisted environments no longer seem like a far dream. We are currently witnessing their slow but seamless integration in everyday lives both at home and at school. Robots have a great potential in being employed as an educational technology. They can be used to facilitate learning and improve educational performance of students in various fields such as physics and mathematics. Independent on their roles in the learning process (passive as teaching aid, co-learner, co-tutors, etc) the main concern is to guarantee their safety around humans. For this, the actions of the robot must be socially acceptable and the results of the actions should be as close as possible to the desired outcome. Several key areas of interest in this respect are movements, environment recognition and kinematics planning for the interaction with the environment. The presented work focuses on robotic manipulation, mainly the automatic identification and evaluation of grasping positions for a set of common objects. The main protagonist in our scenarios is the TIAGo robot produced by the Spanish robot maker PAL Robotics who is having both a passive role as teaching aid and an active one as co-tutors.

This paper addresses the problem of automatic data collection for the purpose of indoor positioning via Received Signal Strength (RSS) fingerprinting. A robotic platform with basic odometer sensors was used in an university building to automate the process of data acquisition which becomes particularly time consuming when considering mapping of large spaces such as shopping malls or hospitals. More than 3000 observations were collected. We associated for each observation their two dimensional coordinates with the received MAC RSS vector. Preprocessing methods included data augmentation and feature normalization. We searched for multiple models and one of the best performance was achieved by using neural networks and post-filtering.

This paper presents a two layer convolutional neural network for performing activity recognition. We combine spatial and temporal information extracted from images acquired from RGB cameras. Spatial information are extracted from videos by splitting them into RGB channel frames and do a one frame at a time classification. Temporal information from videos are extracted by computing their optical flow. The results are combined in order to build a real time human activity recognition system. The network is tested using TIAGo robot for performing activity recognition. The accuracy of the system is 87,05 %, that is comparable with the state of the art. Also, results are obtaining in real time.

For elderly people that are living alone in their homes there is a need to permanently monitor them. One of this aspect consist in knowing their indoor position and motion behavioural status, in real time. One possibility for indoor positioning of an user consists in understanding the images provided by supervising cameras. In this case the main aspect is represented by recognition of objects from these images. Thus, object recognition plays an essential part in understanding the environment and adding meaning to it. This paper presents a method for indoor localisation based on identifying the user’s context. The user’s context is computed based on object recognition and using a probabilistic ontology. The key element is represented by the probabilistic ontology that describes objects, scenes and relations between them. This ontology contains probabilistic relations that are learned using a large database. Results show that given a set of object detectors with high detection rate and low false positive rate, the system can recognize the user’s context with high accuracy.

Physical activity represents a key element for elderly people to maintain a healthy life. This paper presents a game for supporting and stimulating elderly people in performing physical activity. The game is composed of different type of exercises that are selected based on the user's profile and health status. The results are presented to the user in multiple ways that are correlated with his/her emotional state and preferences in order to increase the satisfaction of each user experience.

Magnetic yarns are composite yarns, i.e. they combine elements of various natures and properties, with proven potential for electromagnetic interference (EMI) shielding. In this paper, different mixtures of hard and soft magnetic powder were chosen to cover materials made of cotton yarn. The physical properties and bending behavior of the produced composite yarns were investigated in order to evaluate the yarns for further textile processing.The cotton yarn used as base material was covered with hard (barium hexaferrite BaFe12O19) and soft (Black Toner) magnetic particles. An in-house developed laboratory equipment has been used to cover the twist cotton yarns with seven mixtures having different amounts of magnetic powder (30% – 50%). The bending behavior of the coated yarns was evaluated based on the average width of cracks which appeared on the yarn surface after repeated flexural tests. The obtained results revealed that usage of a polyurethane adhesive in the coating solution prevents crack formation on the surface of hard magnetic yarns after flexural tests. At the same time, the higher the mass percentage of hard magnetic powder in the mixture, the higher was the cracks’ width. The soft magnetic yarns are more flexible and a smaller crack width is observed on their surface. Both the coating solution composition and the powder diameter are expected to influence the bending behavior of coated yarns.

Ambient Assistive Living (AAL) applications allow elderly people to maintain a healthy life style and live longer in their homes. In this paper we describe a platform that combine physical exercises, health monitoring with a reminder component implemented as a multimodal interface adapted to elderly needs for maintaining a healthy lifestyle for elderly people.

Activity recognition plays a key role in providing activity assistance and care for users in intelligent homes. This paper presents a two layer of convolutional neural networks to perform human action recognition using images provided by multiple cameras. We consider one PTZ camera and multiple Kinects in order to offer continuity over the users movement. The drawbacks of using only one type of sensor is minimized. For example, field of view provided by Kinect sensor is not wide enough to cover the entire room. Also, the PTZ camera is not able to detect and track a person in case of different situations, such as the person is sitting or it is under the camera. Also the system will identify abnormalities that can appear in sequences of performed daily activities. The system is tested in Ambient Intelligence Laboratory (AmI-Lab) at the University Politehnica of Bucharest.

In this paper, we present the outcomes and conclusions obtained by involving 220 seniors from three countries (Denmark, Poland and Romania) in a project funded under the European Ambient Assisted Living (ALL) program. CAMI stands for "Companion with Autonomously Mobile Interface" in "Artificially intelligent ecosystem for self-management and sustainable quality of life in AAL". The solution is designed as an innovative architecture that allows for individualized, intelligent self-management which can be tailored to an individual's preferences, culture, level of comprehension, skill, educational needs and learning style. In order to achieve its goals the project has adopted a user-centered approach in which older adults and seniors (CAMI end-users) are involved throughout the project lifetime.

Indoor positioning is one of the major topics in today's navigation and positioning research fields which is partially solved. The research community has not converged to a single, widely accepted solution that can achieve the desired accuracy at the required cost. Wireless Local Area Network (WLAN) based fingerprinting using Received Signal Strengths (RSS) is been considered as one solution for indoor positioning. This study structures this approach as supervised machine learning problem type where the target variable is the position and the features are the RSS values. There are compared the results obtained by from two analysis perspectives, regression and classification.

Regular physical exercises are widely considered to be a key factor for living a healthy life. In this paper we present Mobile@Old, an integrated platform for assisting elderly people to maintain a healthy lifestyle in their homes. Our aim is to highlight the main concepts, technologies, and findings this system rests on. To this end we integrate Mobile@Old in the general conceptual framework of serious games. We provide details about the designing and implementation of Vital Signs Monitoring (VSM) and Physical Activity Trainer (PAT) components of Mobile@Old. Relevant exercises and utilization scenarios are also presented in order to emphases the practical applicability of our approach. We evaluate the usability of platform using the System Usability Scale (SUS). Experimental data regarding the accuracy of whole body movements are also presented

With the continuous ageing of the population, the demand for different healthcare services is increasing at a fast pace. At the same time, the number of caregivers is limited and the cost of well-being is increasing. Therefore, there is a recognized need for technologies that assist elderly people in their daily activities and ensure their safety along with their social integration while maintaining a higher degree of independence. One important aspect in the development of such technologies is their easy acceptance by elderly users. Traditional human-machine interfaces have always represented a barrier for the acceptance of new devices by non-technical people in general and by elderly or people with special needs in particular. In this paper, we propose a multimodal interface designed to match the requirements of Ambient Assisted Living (AAL) systems and to fulfill the needs of elderly people who are the main users of this interface. The interface is multilingual, it supports two languages (English and Romanian) and it can process both speech and gesture commands. The interface is developed using HTML5, JavaScript, CSS3 and integrates Google Speech Service along with other services. It is able to adapt itself to any screen size and to work flawlessly across different platforms. The interface was developed and tested within the "Artificial intelligent ecosystem for self-management and sustainable quality of life in AAL" (CAMI) project. Both the interface and the results of laboratory tests are presented in this paper.

In the context of the rapid growth in the number of electrical and electronic devices and accessories that emit electromagnetic energy in different frequency bands we present and characterize here several magnetic functionalized viscose twisted yarns. A 100% viscose twisted staple yarn was covered through an in-house developed process with a polymeric solution containing micrometric sized barium hexaferrite magnetic powder. The in-house developed process allows deposition of micrometric thickness polymeric paste layer on the yarn surface. Barium hexaferrite is a hard magnetic material exhibiting high chemical stability and corrosion resistivity, relatively large saturation and residual magnetization and microwave absorbing properties. Five different percentages of the magnetic powder in the polymer solution were used, i.e. ranging from 15 wt% to 45 wt%. Physical characterization shows a very good adherence between the highly hygroscopic viscose staple fibers and the polymeric solution that contains polyvinyl acetate and polyurethane as binders. SEM images evidenced the fact that the polymeric solution penetrated more than 1/3 of the yarn diameter. The concentration of magnetic powder in the polymeric solution has a direct influence on the coating amount, diameter and density. The mechanical characterization of the coated yarns revealed that the breaking force is increasing with increasing magnetic powder content up to o certain value and then decreased because the magnetic layer became stiffer. At the same time, the elongation at brake is decreasing.