Program for 2019 IEEE 23rd International Symposium on Consumer Technologies (ISCT)

Time Sala Boxe Aula Didattica Sala Polveri Poster Area Auditorium Break Area Registration Area Outside

Wednesday, June 19

09:00-18:10             Registration  
10:45-11:15           Coffee Break  
11:15-11:45         Open Cerimony    
11:45-12:30         Keynote: Akihiko Sugiyama    
12:30-13:45           Lunch  
13:50-15:30 Consumer Healthcare 1 AI and Deep Learning Applications 1          
15:35-16:20         Keynote: Saraju P. Mohanty    
16:20-16:50           Coffee Break  
16:50-18:10 IoT 1 UI & Multimedia 1          

Thursday, June 20

08:30-17:40             Registration  
09:00-10:20 IoT 2 UI & Multimedia 2          
10:25-11:10         Keynote: Paolo Dario    
11:10-11:40           Coffee Break  
11:40-13:00 AI and Deep Learning Applications 2            
13:00-14:30           Lunch  
14:30-16:10 Consumer Healthcare 2 Energy Management and Security Consumer Electronics for Motion Analysis 1        
16:10-16:40           Coffee Break  
20:00-22:59               Gala Dinner

Friday, June 21

08:30-11:10             Registration  
09:00-10:20 Consumer Healthcare 3 Consumer Electronics for Motion Analysis 2          
10:20-11:10       Poster Session   Coffee Break  
11:10-12:30 Automotive IoT 3            
12:35-13:15 Closing and Award              
13:15-14:30           Lunch    

Wednesday, June 19

Wednesday, June 19 9:00 - 18:10

Registrationgo to top

Room: Registration Area

Wednesday, June 19 10:45 - 11:15

Coffee Breakgo to top

Room: Break Area

Wednesday, June 19 11:15 - 11:45

Open Cerimonygo to top

Room: Auditorium

Wednesday, June 19 11:45 - 12:30

Keynote: Akihiko Sugiyamago to top

Signal Processing for Robot Audition: How Signal Processing Helps a Robot Recognize User Commands in Adverse Environment
Akihiko Sugiyama
Room: Auditorium

This lecture presents an overview of signal processing for robot audition. Human-robot communication is essentially supported by speech recognition whose performance is known to be seriously degraded in adverse environment. To help a robot recognize commands given by the user, four signal processing techniques are useful, namely, direction-of-arrival (DOA) estimation, noise cancellation, echo cancellation, and beam forming. Problems in these techniques which are specific to human-robot communication are identified and solutions to those problems are presented. Video demonstrations in the talk will help audience understand effects of these techniques.

Wednesday, June 19 12:30 - 13:45

Lunchgo to top

Room: Break Area

Wednesday, June 19 13:50 - 15:30

Consumer Healthcare 1go to top

Room: Sala Boxe
Chair: Henry Griffith (Texas State University, USA)
Self-Monitoring of Cardiac Risk while Running Around Ancona
Laura Burattini (Polytechnic University of Marche, Italy); Agnese Sbrollini, Giulia Caraceni, Amnah Nasim, Ilaria Marcantoni, Micaela Morettini, Alberto Belli and Paola Pierleoni (Università Politecnica delle Marche, Italy)
Running is the most common physical activity. Being an aerobic activity, it can act as a trigger for critical cardiac events that may degenerate in sport-related sudden cardiac death. Nowadays, smartphone applications combined with wearable sensors are typically used to monitor runners performance during training, but almost never to evaluate their cardiac risk conditions. Thus, aim of this study was to propose CaRiSMA as a useful tool for self-monitoring of cardiac activity of runners while wearing a cardiac sensor and running by strictly following a route around the city of Ancona (6.1 Km). Cardiac data from10 young runners were recorded and transferred to a smartphone to be analyzed by CaRiSMA, an application that provides two traffic lights as output, relative to cardiac health status of the runner and correctness of training intensity. The first traffic light was green in all cases but one for which it was yellow, indicating no risk and increased risk conditions, respectively. The second traffic light was yellow in all cases, suggesting a reduction of the training intensity. In conclusion, CaRiSMA demonstrated to be a potentially useful tool for self-monitoring of cardiac activity of runners while wearing a cardiac sensor.
NeuroExam: a tool for neurological examination in neuromuscular diseases
Raffaele Conte, Alessandro Tonacci, Francesco Sansone, Maria Scudellari, Anna Paola Pala, Andrea Grande and Gianluca Diodato (Institute of Clinical Physiology (IFC) National Research Council of Italy (CNR), Italy); Marco Calderisi, Francesca Giorgolo and Ilaria Ceppa (Kode, s. r. l., Italy); Stefano Roccella and Andrea Spezzaneve (Scuola Superiore Sant'Anna, Italy); Guja Astrea, Anna Rubegni, Silvia Frosini, Giulia Bertocci and Filippo Santorelli (IRCCS Stella Maris Foundation, Italy)
Neuromuscular conditions are characterized by muscular weakness, for which the subjective clinical phenotype is normally characterized by means of neurological examination. However, this assessment is dramatically affected by inter-observer variability, possibly impacting on the patients' diagnosis. To help solving this issue, an interoperable HW/SW solution was implemented and preliminarily tested in a clinical setting, representing a useful tool for collecting data in a safe, secure and reliable way, and providing the clinician with a valid tool for objectivizing the parameters acquired during the neurological exam. Thanks to the data elaboration part, the solution can undoubtedly represent a valid tool to support the diagnostic path and can potentially be applied to several conditions, not only in the neuromuscular domain but also in the whole neurological field.
Consumer Adoption of Digital Technologies for Lifestyle Monitoring
Luca Gastaldi, Emanuele Lettieri and Marco Mandolfo (Politecnico di Milano, Italy)
Despite their potential, the adoption of wearable devices has been relatively slow when compared to other digital technologies. This paper investigates, grounding on the Theory of Planned Behavior, the adoption by end users of digital technologies for lifestyle monitoring. Data on consumers' perception and usage of wearable devices have been collected through a survey administered to 1,000 Italian citizens and further analyzed through a Structural Equation Model approach. Results show that, above the functional value of the device, external influence, particularly doctor opinion, exerts an essential role in adoption. Online health literacy proves to be a relevant factor as well, showing the importance of cultural patterns in wearables diffusion. Implications for academicians, practitioners and policy-makers are provided.
The Role of Mobile Apps in Heart Rate Measurement with Consumer Devices
Susanna Spinsante (Università Politecnica delle Marche, Italy); Valeria Cerquetti and Angelica Poli (Universita' Politecnica delle Marche, Italy); Lorenzo Scalise (Università Politecnica delle Marche, Italy)
Consumer mobile health technologies offer potentially disruptive opportunities to the different stakeholders involved in the healthcare process, but further investigations concerning the accuracy and reliability of physiological parameters measurement by means of wearable devices need to be performed. For example, experiments show that under the same conditions, different mobile apps may return different heart rate values, at a parity of the smartwatch used, so this work addresses the role of the mobile apps within the measurement process, to understand if differences exist, and if they may significantly affect the results provided to the user. Tests executed with the same smartwatch and different commercially available apps, at rest and during walking, show that there is no relevant difference in the heart rate measurements, but further investigations are needed to confirm the validity of this preliminary outcome.
Smartwatch based emotion recognition in Parkinson's disease
Lucia Pepa, Marianna Capecci and Maria Gabriella Ceravolo (Università Politecnica delle Marche, Italy)
Parkinson's disease (PD) is a neurodegenerative disorder characterized by several motor and non-motor symptoms. These last, widely, concern affective disorders, among which, alexithymia has great impact on the amount and quality of social relationship and on caregiver well-being. Ubiquitous and quantitative monitoring of patients' emotional status may help to better understand affective disorders. This study proposes a preliminary approach to investigate the feasibility of emotion recognition in PD using physiological data collected by a smartwatch.

Wednesday, June 19 13:50 - 15:10

AI and Deep Learning Applications 1go to top

Room: Aula Didattica
Chair: Francesco Ferracuti (Università Politecnica delle Marche, Italy)
Evaluation of Deep Convolutional Neural Network architectures for Emotion Recognition in the Wild
Abudukaiyoumu Talipu, Andrea Generosi and Maura Mengoni (Università Politecnica delle Marche, Italy); Luca Giraldi (Emoj SRL, Italy)
This paper presents a software based on an innovative Convolutional Neural Network model to recognize the six Ekman's universal emotions from the photos of human faces captured in the wild. The CNN was trained using three different datasets already labeled and merged after making them homogeneous. A comparison among different types of CNN architectures using the Keras framework for Python language is proposed and the evaluation results are presented.
Epileptic seizures prediction based on the combination of EEG and ECG for the application in a wearable device
Lucia Billeci and Alessandro Tonacci (Institute of Clinical Physiology (IFC) National Research Council of Italy (CNR), Italy); Maurizio Varanini (Institute of Clinical Physiology (IFC), National Research Council of Italy (CNR), Italy); Paolo Detti, Garazi Zabalo Manrique De Lara and Giampaolo Vatti (University of Siena, Italy)
Epilepsy is a neurological disorder characterized by recurrent and sudden seizures. Recently, researchers found that patients often present physiological abnormalities that precede an epileptic seizure onset. Importantly, these modifications can vary a lot among patients. While the conventional methodology for characterizing epilepsy is electroencephalogram (EEG), some evidences show that electrocardiogram (ECG) can be also useful to assess modifications associated to seizures. In this paper, a preliminary study about the integration of EEG and ECG for a patient-specific seizure prediction is presented. Synchronization patterns from the EEG and time and frequency features, as well as recurrence quantification analysis measures from the RR series, were extracted. A support vector machine (SVM) classifier was then applied to classify preictal and interictal phases combining features extracted from the two signals. Results showed that, using the proposed combined approach it is possible to predict the epileptic seizure onset with a total average sensitivity of 93.3%, specificity of 80.6% and a prediction time of about 20 min. This approach could be implemented in portable and wearable devices for a real-life patient-specific seizure prediction.
Crowd emotion detection to light up a smart Christmas tree
Massimo Martini, Andrea Felicetti, Marco Mameli and Raffaele Vaira (Università Politecnica delle Marche, Italy); Salvatore La Porta and Fabrizio Marconi (JEF Srl, Italy); Isabella Lazzini (Huawei Consumer Business Group Italia, Italy); Rocco Pietrini (Università Politecnica delle Marche, Italy)
The paper aims to describe part of a system created by Huawei, that proposes the first Christmas tree endowed with artificial intelligence. It is able to recognize facial emotions from images acquired by a mobile application and to light up itself with different lights and colors according to the prevalent sentiment. In this project, our role was to test the neural network integrated in the mobile application and able to recognize the sentiment of a facial expression. We have implemented a model based on a convolutional neural network and tested the accuracy of the recognition method, by creating an "ad-hoc" dataset of images.
Towards the Design of a Machine Learning-based Consumer Healthcare Platform powered by Electronic Health Records and measurement of Lifestyle through Smartphone Data
Alessandro Ferri and Riccardo Rosati (Università Politecnica delle Marche, Italy); Michele Bernardini (Università Politecnica delle Marchei, Italy); Leonardo Gabrielli and Sara Casaccia (Università Politecnica delle Marche, Italy); Luca Romeo (Universita' Politecnica delle Marche, Italy); Andrea Monteriù and Emanuele Frontoni (Università Politecnica delle Marche, Italy)
The estimation of Biological Age (BA) has been debated for several years and no clear and universal understanding has yet been reached to solve this task. Accordingly, the knowledge of an accurate BA index for each individual may be relevant in various areas including health, economy, social policies and decision making processes. The main contribution of this work is the design of a Machine Learning based-consumer healthcare platform powered by electronic health record data (clinical features) and smartphone data (lifestyle features) in order to estimate a sub-index that is strictly correlated with the BA. Preliminary results extracted from a representative subset of clinical and lifestyle features, highlight the potential of the proposed framework in order to estimate the health and physical status of each subject (in terms of the difference between the predicted Chronological Age and the real Chronological Age). Future work will be conducted to encapsulate more information and validate the predicted BA sub-index.

Wednesday, June 19 15:35 - 16:20

Keynote: Saraju P. Mohantygo to top

Consumer Technologies for Smart Cities
Saraju P. Mohanty
Room: Auditorium

The smart cities have been envisioned to mitigate the problems of rapid migration of human population in in both man-made and natural resources constraint. The smart cities use one or multiple smart systems including smart healthcare, smart transportation, smart agriculture, smart infrastructure, and smart grids, and hence in an essence is a system of systems. The systems of the smart cities are essentially cyber-physical systems or CPS which are built using Internet of Things (IoT). IoT is a configurable dynamic global network of networks has components like: The Things, Internet, LAN, and The Cloud. The IoT infrastructure consists of various elements including sensors, electronics, networks, middleware, firmware, and software, which are essentially consumer technologies. The objective of this talk is to analyze the consume technologies that built smart cities. In this keynote, the various components of the smart cities and the underneath consumer technologies will be elaborated. Specific technologies like Sensors for diverse smart cities applications, Unmanned Arial Vehicle (UAVs), Camera Technology, Blockchain, Physical Unclonable Functions (PUF), and Artificial Intelligence (AI) in the context of smart cities will be discussed. The audience will find answers to several questions on smart cities and corresponding consumer technologies, including the following: (1) What is a smart city? (2) What are the important components of smart cities? (3) What makes systems smart? (4) What are critical consumer technologies for smart cities? (5) What are the challenges of smart cities? (6) What are the research directions for the design and operation of efficient smart cities?

Wednesday, June 19 16:20 - 16:50

Coffee Breakgo to top

Room: Break Area

Wednesday, June 19 16:50 - 18:10

IoT 1go to top

Room: Sala Boxe
Chair: Toru Kobayashi (Nagasaki University, Japan)
Feasibility of Recommendation System Mediated by Social Internet of Vehicles
Arshad Muhammad (Sohar University, Oman); Kashif Zia (Sohar University & Islamabad, Oman)
The latest manifestation of all connected world is the Internet of Things (IoT), and Internet of Vehicles (IoV) is evolved as one of the first prominent examples of IoT. Social IoV (SIoV) is a term used to represent when vehicles build and manage their own social network. While exploring these futuristic systems, in addition to physical aspects, the social aspects of connectivity and information dispersion should also be considered. In this paper, an agent-based model of information sharing (for context-based consumer recommendations) of a hypothetical population of smart vehicles is presented, in order to check the feasibility to offer recommendation system for the consumers. The simulation results show that the closure of social ties and its timing influences the dispersion of novel information considerably. It is also observed that as the network evolves as a result of incremental interactions, the recommendations guaranteeing a fair distribution of vehicles (consumer-owned cars) across equally good competitors is possible.
CPS/IoT Ecosystem: Indoor Vertical Farming System
Haris Isakovic (Technische Universität Wien, Austria); Radu Grosu (Vienna University of Technology, United Kingdom (Great Britain)); Lukas Punzenberger (Technsiche Universität Wien, Austria); Alexander Fasching (Technische Universität Wien, Austria)
Building large-scale IoT applications requires enormous infrastructural support. As infrastructure we understand hardware, software and communication channels necessary to ensure interconnection between various heterogeneous components. Project CPS/IoT Ecosystems explores infrastructural requirements for large-scale applications. The applications are real-world scenarios such as smart parking, smart agriculture or smart buildings. In this paper we explore infrastructural requirements necessary to build modular indoor vertical farming system. The proposed prototype is a service-oriented platform distributed over three scopes of operation: cloud, fog, sensor/actuator.
Designing, implementing and testing an IoT based home system for integrated care services
The purpose of this paper is to describe the design, development and implementation plan for testing the transferability of an ICT based solution for Integrated Care (the ProACT system), originally optimized for older people with multimorbidity, to an enlarged group of users and needs. The transfer trial will be held in Italy and requires the development of a specific expansion of the ICT platform and service protocols. This paper outlines the details of the design, development and implementation plan, including the home IoT based solution. The target participants will be disabled people, informal carers, formal carers, professionals in the health and social domain.
IoT Solution based on MQTT Protocol for Real-Time Building Monitoring
This paper presents an IoT architecture for continuous and real-time building monitoring. The proposed solution is based on a sensor node positioned inside the building to be monitored, connected to the Internet, able to perform continuous measurements and to real-time send raw data to a remote server through the MQTT protocol. Thanks to the development of a specific Web application, each authorized user can access the real-time information on the health status of the structure, the history of the measurements and, if necessary, send commands to the sensor node. The goal is to provide a low-cost, reliable and replicable system for the implementation of a widespread sensors network for structural monitoring.

UI & Multimedia 1go to top

Room: Aula Didattica
Chair: Bill Kapralos (University of Ontario Institute of Technology, Canada)
Re-design of the Household Appliance UI to make it an Adaptive System
Francesca Gullà, Alessandra Papetti, Agnese Brunzini and Michele Germani (Università Politecnica delle Marche, Italy)
Everyday life is increasingly rich in man-machine interactions and new challenges in user interface design arise. In particular, it emerges the need of adaptable solutions that learn from the users behavior to improve their experience. In this context, the paper aims to redesign an existing UI to make it an Adaptive System. The introduction of an adaptive module allows finding the optimal interface features combination based on the user profile and previously interactions. The experimentation results demonstrate the adaptability and versatility of the proposed application by evaluating the user satisfaction and the perceived adaptability with respect to the native application.
Enabling by Voice
Umber Shamim (Brunel University, United Kingdom (Great Britain)); Gabriella Spinelli (Brunel University London, United Kingdom (Great Britain)); Alan Woodcock and Ajoy Nair (The Hillingdon Hospital Trust, United Kingdom (Great Britain))
Interactive Smart Agents (ISA) like Amazon Echo are rapidly becoming part of our households. These devices are controlled by the user through conversations in natural language. In turn, these agents control other smart devices around the house. Similar control functionalities are provided by Environmental Control (EC) devices issued to people with severe mobility impairments. EC devices are controlled by scan and click methods (also known as switch methods) such as eye gaze, 'suck and puff', etc. [1]. EC devices are prescribed by the NHS through regional EC services like the North Thames Regional Environmental Control Equipment Service (NTRECES). During the year 2016-17 NTRECES had 86 Spinal Cord Injury (SCI) clients out of the total 732 clients they look after. In severe cases of SCI people suffer complete loss of mobility that affects them from the neck down. However typically, they retain excellent cognitive and communication skills. In such cases, patients have to use their head to click the switch, use suck and puff device or IntegraMouse (controlled by lips) to sequentially go through all the menu options. These methods are time consuming, frustrating and undignified [2]. Currently the available input mechanisms rely on the little available mobility which these patients have left to access and don't allow them to use the excellent cognitive and communication skills they still retain. This paper argues that an EC device based on voice-controlled ISA could be an option to enable SCI patients to have a better quality of life. The current design and development of ISA devices makes them dependent on internet for all the speech processing. This means that in cases of internet or power failure ISA devices have no backup, providing no options for the users they support. In addition, due to the reliance of the ISA devices on the internet, it makes them more vulnerable to security and privacy breaches. Moreover, the accuracy of speech recognition is affected by the presence of background noise or speech impairments, which makes the devices suitable for patients that have full and unhindered speech. ISA devices are designed as mainstream consumer technology. The patients' demand for ISA instead of traditional EC devices is growing. However due to the shortfall describes ISA are only prescribed in pilot cases. If the technology and design of ISA were to be addressed, they would represent a significant saving to the national health services. This paper aims to highlight the need for a user-centred investigation on the design of ISA based EC devices.
A Dialogue Intervention Simulation Framework to Facilitate Psychotherapy Training
Kyle Wilcocks (UOIT, Canada); Bill Kapralos (University of Ontario Institute of Technology, Canada); Allen Kalpin (IEDTA, Canada); Allan Abbass (Centre for Emotions & Health, Dalhousie University, Halifax, Nova Scotia, Canada); David Chandross (G. Raymond Chang School of Continuing Education, Ryerson University, Toronto, Canada)
Intensive Short-Term Dynamic Psychotherapy (ISTDP) is a form of psychotherapy that evokes emotional experiences with the patient to facilitate a corrective emotional experience. During ISDTP, verbal and non-verbal communication skills play an essential role in how the dialogue is directed to achieve successful treatment. In an attempt to enhance non-verbal communication skills, we present a framework that allows for the simple creation of a dialogue between the user and a virtual avatar representing the patient. The patient avatar incorporates facial expressions to express specific non-verbal communication thus expose psychotherapy trainees to better simulated non- verbal cues.
User Perception of Robot's Role in Floor Projection-based Mixed-Reality Robotic Games
Within the emerging research area represented by robotic gaming and, specifically, in application domains in which the recent literature suggests to combine commercial off-the-shelf (COTS) robots and projected mixed reality (MR) technology in order to develop engaging games, one of the crucial issues to consider in the design process is how to make the player perceive the robot as having a key role, i.e., to valorize its presence from the user experience point of view. By moving from this consideration, this paper reports efforts that are being carried out with the aim to investigate the impact of diverse game design choices in the above perspective, while at the same time extracting preliminary insights that can be exploited to orient further research in the field of MR-based robotic gaming and related scenarios.

Thursday, June 20

Thursday, June 20 8:30 - 17:40

Registrationgo to top

Room: Registration Area

Thursday, June 20 9:00 - 10:20

IoT 2go to top

Room: Sala Boxe
Chair: Ezio Bartocci (Vienna University of Technology, Austria)
RPA Constitution Model for Consumer Service System based on IoT
Toru Kobayashi, Kenichi Arai and Tetsuo Imai (Nagasaki University, Japan); Tomoki Watanabe (NTT Corporation, Japan)
Currently, RPA (Robotic Process Automation) attracts attention for productivity improvement of the business processing. However, there are few examples that applied RPA to consumer service system. It is caused by the fact that there is not common sense about an application development method based on RPA for consumer service system. Therefore, we suggest a basic model of RPA constitution for consumer service system. The feature is that we can add continuity, automation, and usability to consumer service system. We inspect an effectiveness of the basic model of RPA constitution by using consumer service system examples characterized by fusing IoT and AI that we worked on so far.
Approximate Matrix Inversion Methods for Massive MIMO Detectors
Mahmoud A. M. Albreem (A'Sharqiyah University (ASU), Oman)
Building a practical detector to achieve the attractive merits of large-scale multiple-input multiple-output (MIMO) is not a trivial task. However, low-complexity linear detectors could achieve a satisfactory performance for uplink massive MIMO, but such detectors involve unfavorable matrix inversions, whose not a hardware friendly. This paper considers several approximate matrix inversion methods to avoid a direct matrix inversion, namely, the Neumann method, the Gauss-Seidel (GS), the successive over-relaxation (SOR), the Jacobi method, the Richardson method, the optimized coordinate descent (OCD), and the conjugate gradients (CG) method. This paper shows that the MMSE detector based on the SOR and GS methods outperform the other detectors when the ratio between the number of BS antennas and user terminal antennas (β) varies from small to large values. It also investigates the selection of ω for both Richardson and SOR methods. The optimum performance can be achieved when ω = 2 λ in the Richardson method. This paper shows that not every iteration has a positive impact on the performance.
Smart Aquaculture System: A Remote Feeding System with Smartphones
Tetsuo Imai, Kenichi Arai and Toru Kobayashi (Nagasaki University, Japan)
The problem of open water aquaculture using submersible fish cages is to reduce costs of labor and feed. In this paper, we propose the Smart Aquaculture System to reduce labor and feed costs by the labor saving of aquaculture, including automated feeding and continuous observation of cages, optimization of a timing and an amount of feeding. As a part of the system, we implemented the Remote Feeding System comprised of a feeding device equipped inside the cage and a smartphone to control the device remotely. We confirmed the proper operation through field experiments including underwater.
Capacitive Soil Moisture Sensor Node for IoT in Agriculture and Home
Christian Hirsch (Technische Universität Wien, Austria); Ezio Bartocci (Vienna University of Technology, Austria); Radu Grosu (Vienna University of Technology, United Kingdom (Great Britain))
Growing food for nine billion people is a challenge we will face in near future. Digital technologies hold a great promise to increase efficiency and use resources effectively. In this paper we present a low-power and scalable IoT-based architecture for home farmers and scientific purposes that enables to verify the environmental impact on plants developments by monitoring the soil moisture and temperature. The measured values are transmitted via Bluetooth Low Energy (BLE) to a gateway, e.g., a smartphone running the gateway app, to an cloud platform that stores and processes the data. This data is useful to give home farmers vital information, e.g., when to water a plant or when the plant is prune to get infected by a disease. This information is of great importance because it helps to decrease crop failures and thus farm more efficient. The hardware and software architecture presented is scalable and designed to be consume low power. Experiments show that it is possible to create a high spatial resolution by putting many sensor nodes on a small area in order to help research to measure micro-climate.

UI & Multimedia 2go to top

Room: Aula Didattica
Chair: Wen-Chung Kao (National Taiwan Normal University, Taiwan)
Development of a Virtual Reality-Based Working at Heights Safety Awareness Framework
Kyle Wilcocks (UOIT, Canada); Andrei Torres (UOIT, Canada & Universidade Federal do Ceara, Brazil); Alvaro Uribe Quevedo and Bill Kapralos (University of Ontario Institute of Technology, Canada); Shannon Hunt, Kim Slade and Glenn Cullen (Public Services Health and Safety, Toronto, ON, Canada)
Working at heights leads to serious injuries and deaths globally each year in part, due to poor working conditions, improper use of protective equipment, and lack of training. Currently, safety is enforced by requiring mandatory working at heights certification whereby trainees take written and practical tests in regular classrooms or specialized facilities (if available), to become licensed/certified. However, despite training, certification, and policy efforts, a number of accidents are still occurring when working at heights. In this paper, we present a virtual reality- based simulation that safely exposes and immerses trainees to more realistic training at heights scenarios. The simulation also includes a practical test component that involves 100% tie-off travel restraint on top of a building.
Power Modeling for Virtual Reality Video Playback Applications
Christian Herglotz (Friedrich-Alexander University Erlangen-Nürnberg (FAU), Germany); Stéphane Coulombe (École de Technologie Supérieure, Canada); Ahmad Vakili (Summit-Tech, Canada); Andre Kaup (University of Erlangen-Nuremberg, Germany)
This paper proposes a method to evaluate and model the power consumption of modern virtual reality playback and streaming applications on smartphones. Due to the high computational complexity of the virtual reality processing toolchain, the corresponding power consumption is very high, which reduces operating times of battery-powered devices. To tackle this problem, we analyze the power consumption in detail by performing power measurements. Furthermore, we construct a model to estimate the true power consumption with a mean error of less than 3.5%. The model can be used to save power at critical battery levels by changing the streaming video parameters. Particularly, the results show that the power consumption is significantly reduced by decreasing the input video resolution.
Head Motion Compensation for Visible-Spectrum Gaze Trackers
Wen-Chung Kao, Shao-Ching Lin, Jyun-Yi Li, Yi-Chin Chiu and Kuan-Jen Huang (National Taiwan Normal University, Taiwan)
The visible-spectrum gaze tracker (VSGT), which is designed without the extra infra-ray (IR) illuminations, gains the advantage of better user experiences than the traditional IR-based gaze tracker. However, the design of head motion compensation for a VSGT becomes more difficult due to the lack of stable glints on the human eyes. In this paper, a new head motion compensation mechanism for the binocular VSGT has been proposed. The improvement comes from the fact that the proposed approach accurately estimates the relative positions/poses among the camera, the screen, and the human eyes via the proposed two-phase calibration. The experimental results show that the proposed approach can significantly improve compensation result for the wide range of head movement.
Interactive 360° Video and Storytelling Tool
Louay Bassbouss and Stephan Steglich (Fraunhofer FOKUS, Germany); Igor Fritzsch (BitTubes GmbH, Germany)
Existing 360° video players on the market playback only one type of media in their timeline. In this paper, we will introduce a new tool that allows creating interactive 360° videos with storytelling elements that can switch between flat and 360° videos in one timeline. The tool also integrates other media types like images, audio and web resources that can be interlinked with each other to create one or multiple stories. The benefits of this approach are far-reaching i.e. it allows broadcasters to create genres of programming that combine both traditional footage and immersive 360° video segments in one seamless experience. This enables a highly effective and attractive lean-back and lean-in experience for the audience. Moments of immersion can then be coupled with traditional storytelling, which is an attractive option for content creators who can enhance existing material with immersive moments in a way that is cost effective and allows them to fall back on their pre-existing skills as story and programme creators. This will pave the way for much broader adoption and production of spherical content among content providers of the first wave, as it lowers the barrier of entry. The tool will allow content producers and editors to create stories in 360° videos and in combination with non-360° media content by providing an easy-to-use web-based editor that targets devices with various characteristics and input capabilities such as TVs, tablets and head-mounted displays (HMDs).

Thursday, June 20 10:25 - 11:10

Keynote: Paolo Dariogo to top

Robotics in the future
Paolo Dario
Room: Auditorium
Chair: Sauro Longhi (Università Politecnica delle Marche, Italy)

Thursday, June 20 11:10 - 11:40

Coffee Breakgo to top

Room: Break Area

Thursday, June 20 11:40 - 13:00

AI and Deep Learning Applications 2go to top

Room: Sala Boxe
Chair: Gordana Velikić (RTRK Computer Based Systems LLC, Serbia)
Automatic detection of canonical image orientation by convolutional neural networks
Lia Morra, Sina Famouri, Huseyin Karakus and Fabrizio Lamberti (Politecnico di Torino, Italy)
Automatically detecting the canonical orientation of images based on their visual content is an essential part of many computer vision and image processing pipeline. It allows to disregard meta-data such as EXIF which may not be consistently stored or manipulated across different devices, applications, and file formats. Similar to human perception, convolutional neural networks can exploit implicit object recognition and other useful semantic cues to predict the correct orientation of an image. In this work, we leverage the properties of convolutional neural networks and adapt a pre-trained model to the image orientation detection task. It is shown by extensive evaluation that our method can work well on a range of datasets, including relatively low quality photos generated by a wide range of consumer devices. On public benchmarks, our method compares favourably with the state-of-the-art and achieves accuracy very close to that of humans.
Development of a Mimic Robot: Learning from Human Demonstration to Manipulate a Coffee Maker as an Example
Chen-Chien Hsu, Pin-Jui Hwang and Wei-Yen Wabg (National Taiwan Normal University, Taiwan)
With the trends of DIY movements and the maker economy, great needs for the applications of low volume automation (LVA) are predictable. To do so, a learning from demonstration (LfD) problem is addressed in our research, where robots are taught through demonstrated actions to manipulate a coffee maker. The system is developed using YOLO deep learning architecture to recognize objects such as cups, coffee capsules. Employing a Kinect RGB-D camera, the robot is capable of obtaining the coordinates of the objects and the corresponding moving trajectories. Integrating the above two techniques, the robot is able to recognize the demonstrated actions to establish an action database comprising several sub-actions such as moving a cup, pouring coffee, trigger the coffee machine, etc. Finally, the manipulation of the robot is made by following the order of demonstrated actions. As a result, a vision-based LfD system is established, allowing the robot to learn from human demonstrations and act accordingly.
Image Compression at Very Low Bitrate Based on Deep Learned Super-Resolution
Sébastien Hamis and Titus Zaharia (Télécom SudParis, France); Olivier Rousseau (Be-Bound, France)
The problem of data storage and transmission on mobile devices is constantly growing up. Smartphones are nearly by default equipped with HD cameras that are taking high quality pictures, which can be instantly stored and easily uploaded on cloud platforms. Such a behavior favors the creation of a massive amount of data. In order to reduce the size of such data, it is mandatory to dispose of efficient compression techniques that can take into account the actual usage of such image data. For example, most of the pictures acquired by phone cameras are often displayed on a small screen and this, for a little amount of time. A solution to manage this kind of oversized data would be to store them in a lower resolution, additionally to a standard image compression. The downside of such an approach is that restoring an image to its original resolution is a challenging task, notably in the presence of complex compression artifacts, such as those introduced by sophisticated compression methods. In order to deal with such an issue, in this paper we propose a new model, specifically trained to perform super-resolution after compression with the BPG state-of-the-art codec. An advantage of the proposed approach comes from the fact that the underlying process can be interpreted as a post-processing step, which cancan be easily added to any compression scheme without modifying the codec. Experimental results show that our model outperforms perceptually the state-of-the-art compression standards even for very low bitrates.
Machine Learning Based App-Collusion Detection in Smartphones
Mohammad Faiz Iqbal Faiz (North Eastern Regional Institute of Science and Technology, India)
The App-collusion is a multi-app attacking model in which two or more apps work together to achieve a malicious goal. Single app analyses tools are not able to detect app-collusion because they focus on single apps. This paper presents a novel and simple technique to detect collusive app-pairs using machine learning classifiers. The method employs a two-stage classification model. In the first stage, we have logistic regression based classifier to detect single malicious applications. In the second stage, we use the parameters obtained from the first stage to detect collusive app-pairs. In the second stage, we use either logistic regression or perceptron model. Our two-stage classification model is successful in detecting colluding app-pairs.

Thursday, June 20 13:00 - 14:30

Lunchgo to top

Room: Break Area

Thursday, June 20 14:30 - 16:10

Consumer Healthcare 2go to top

Room: Sala Boxe
Chair: Federica Verdini (Universita' Politecnica delle Marche, Italy)
HSV Conversion Based Tactile Paving Detection for Developing Walking Support System to Visually Handicapped People
Yuki Ito and Chinthaka Premachandra (Shibaura Institute of Technology, Japan); Sagara Sumathipala (University of Moratuwa, Sri Lanka); Bopagoda Hettiarachchige Sudantha (University of Moratuwa & Shibaura Institute of Technology, Sri Lanka)
In this research, we present a system that aims to support visually handicapped people in walking alone. It warns the user of any danger by detecting the tactile paving and recognizing obstacles on it. Thus, potential accidents involving visually handicapped people can be prevented. The system guides the user by interpreting the information obtained from the automatic detection of the tactile paving. In this paper, the automatic tactile paving is mainly discussed by processing images from on-user camera which exists in the main system. We propose a new HSV conversion-based image processing approach to detect them. Ultimately, the results of the experiments confirm the achievement of high tactile paving detection rates.
TWA Identifier for Cardiac Risk Self-Monitoring during Hemodialysis A Case Report
Chiara Leoni, Ilaria Marcantoni, Agnese Sbrollini and Micaela Morettini (Università Politecnica delle Marche, Italy); Laura Burattini (Polytechnic University of Marche, Italy)
Rate of sudden cardiac death (SCD) is increased in hemodialysis (HD) patients. Cardiac risk can be evaluated in terms of electrocardiographic (ECG) T-wave alternans (TWA). Aim of the present study was to propose TWA Identifier as a software application for cardiac risk self-monitoring based on TWA index, and to test it on a patient while performing a HD session. TWA Identifier can be installed in any portable device and may analyze real time ECG data acquired by wearable sensors. The core algorithm of the TWA Identifier is the well-established heart-rate adaptive match filter for TWA identification. Here it was used to quantify TWA from a continuous 24 hours ECG acquired using a wearable Holter ECG recorder in a HD patient during a HD day. The recording was divided into macro-time periods, one prior, one contemporary and two following the HD session. On average, TWA values were higher than normal, ranged from 35 µV to 78 µV, and were particularly high during the HD session while decreased afterwards. Thus, the HD patient was at increased SCD risk, especially during the treatment. In conclusion, TWA Identifier represents a useful tool for real time cardiac risk self-monitoring during HD.
Container Type and Fill Level Classification Using a Bottle-Attachable IMU Sensor
Henry Griffith (Texas State University, USA); Subir Biswas (Michigan State University, USA)
Attachable inertial measurement unit sensors are a promising non-wearable approach for tracking fluid consumption across multiple drinking vessels. This paper demonstrates the ability of these devices to classify the type of container to which they are attached, along with the initial fill level from which a drink is consumed. Classification is performed using support vector machine models with hand-engineered features describing the container's estimated inclination during drinking. Three common container types (refillable bottle, glass, and mug) and two fills levels are considered herein for preliminary proof-of-concept. Variability in performance versus training strategy (i.e.: subject-specific versus out-of-subject training) and kernel functions is explored. Subject-specific models are demonstrated to classify container type with 98.7% and 88.0% accuracy at the two fill levels considered. Initial fill level is classified with 100% accuracy for each container type using subject-specific models. For an equivalent number of training samples, subject-specific models improve container classification accuracy by nearly 10% versus models trained out-of-subject for the various scenarios considered.
Hand-crafted Features vs Residual Networks for Human Activities Recognition using Accelerometer
Anna Ferrari, Daniela Micucci and Marco Mobilio (University of Milano-Bicocca, Italy); Paolo Napoletano (University of Milan, Bicocca, Italy)
Inertial sensors combined with supervised machine learning techniques are largely employed for automatic Human Activity Recognition (HAR). Machine learning scientists made available to the community a plenty of labeled databases that permit, especially in the recent years, to develop sophisticated techniques, such the ones based on deep learning. These techniques have recently become very popular because they are highly accurate. Nevertheless, some researchers still use the combination of traditional classifiers, such as SVM and k−NN, with hand- crafted features or raw signals. The aim of this paper is to investigate the robustness of traditional classifiers combined with hand-crafted features compared with an end-to-end deep learning solution based on a Residual Network. Experiments on four public databases are presented and discussed.
Recurrence Analysis of Human Body Movements during Activities of Daily Living
Amnah Nasim, Micaela Morettini, Ilaria Marcantoni and Agnese Sbrollini (Università Politecnica delle Marche, Italy); Laura Burattini (Polytechnic University of Marche, Italy)
Recurrence quantification analysis (RQA) is used to differentiate and analyze the regular and irregular parts of a time-series signal using recurrence plots and quantification measures. This work presents RQA for human body movements during routine activities of daily life (ADL) using parameters recorded using a wearable sensor attached to the test subject's waist. The current research uses data from 8 subjects performing 5 different daily life activities, lying and stand, pick and stand, sitting and stand, step up and down, and walking. Simulating the RQA plots for activity and non-activity phases for squared vector magnitude parameter for each of the record we quantify the level of signal stability and disruption in terms of RQA analysis measures recurrence rate (RR), determinism (DET) line entropy (ENT). The RQA parameters reveal a chaotic behavior in case of activity (RR=0.249, DET=0.510, ENT=0.732), and a stable or least chaotic behavior in case of non-activity (RR=0.0.466, DET=0.726, ENT=1.205) regions of time. Distinguishing values for RQA-based measures for different human body movements taking place during daily life activities might be used for human activity monitoring, fall detection for elderly and body movement modelling and analysis algorithms.

Energy Management and Securitygo to top

Room: Aula Didattica
Chair: Jose-Maria Flores-Arias (University of Cordoba, Spain)
An Overview of Blockchain-based Applications for Consumer Electronics
Valentina Gatteschi and Fabrizio Lamberti (Politecnico di Torino, Italy); Claudio Demartini (Polithecnic of Turin, Italy)
A high number of research initiatives investigated the advantages of blockchain technology in different sectors. The objective of this work is to present an overview of blockchain-based applications in the consumer electronics field. In particular this work reports around fifty blockchain initiatives in the contexts of healthcare, smart homes, smart cities, automotive and smart devices, and discusses advantages/limitations of this technology.
A blockchain technology for privacy protection and probative value preservation of vehicle driver data
Claudio Ferretti (Università degli Studi di Milano - Bicocca, Italy); Francesco Morano and Alberto Leporati (Università degli Studi di Milano-Bicocca, Italy); Paolo Napoletano (University of Milan, Bicocca, Italy); Raimondo Schettini (University of Milano Bicocca, Italy)
The number of sensors in cars and other vehicles has significantly increased in the last few years. Such sensors are used to monitor vehicle telemetry as well as driver's biometric and physiological data, with the aim of increasing the safety of people and decreasing the number of accidents. Data monitored by these sensors may also be used by insurance companies or police for accident, or more in general, event reconstruction. Whatever is the use, this data should be protected against cybersecurity attacks in order to guarantee the privacy of the drivers, and to prevent malicious modifications performed with the aim of cheating the insurance companies or police. The goal of this paper is to propose a framework, based on a permissioned blockchain, that allows to both guarantee driver data protection and evidential property of data. The scenario explored considers data recorded inside the car, encrypted and transferred to a cloud storage service. The encrypted version of the data is also hashed and stored in the blockchain, to guarantee data integrity and inalterability. At the same time, the blockchain can be used as an official ledger for preserving the probative value of data.
Feasibility Analysis of several RES installations for a NGO in Ethiopia through a Developed Holistic SW Tool
Antonio Escribano (Universtiy of Cordoba, Spain); Jose-Maria Flores-Arias (University of Cordoba, Spain); Francisco J. Bellido-Outeiriño and Joaquín Garrido (University of Córdoba, Spain); Emilio J. Palacios-Garcia (Aalborg University, Denmark)
Being able to ensure the viability of a renewable energy source (RES) installation is always the first obstacle to be overcome before making any investment. This is especially important when the installation is not going to be made use commercially and is not going to return benefits in the short or long term. This proposal shows the practical application of a SW tool developed for energy, financial and environmental feasibility studies to a demand from non-governmental organizations (NGOs) missions located in Ethiopia to study the possibility of installing photovoltaic generation plants to supply their basic school and dispensary facilities.
Demand Side Flexibility in the Residential Scenario: A survey of techniques
Paolo Falcioni (APPLiA Home Appliance Europe, Belgium); Lenka Jancova (Smart Living and Competitiveness Policy Specialist & APPLiA Home Appliance Europe, Belgium)
Decentralization, decarbonization and digitalization are the main trends transforming the energy system in Europe and beyond. This paper aims to outline a consumer perspective of the changing energy market and new, emerging roles and opportunities it brings along. Consumers will play a vital role in this transformation through adapting their power demand. While it is an overall aim to increase the integration of renewable energy sources and these sources are not available at demand, new technologies and new algorithms need to be put in place to better benefit from them. The paper thus presents a survey of the best DSM residential methodologies from both an academic and industrial perspective. It also indicates flexibility potential of different appliances and obstacles that need to be overcome to materialize this potential.
Decimation Structures for Power of Three Decimation Factors for Consumer Devices
Gordana Jovanovic Dolecek and Carlos Santin Cruz (Institute INAOE, Mexico)
This paper presents a low-power non-recursive decimation structure for power of three decimation factors suitable for consumer devices. The proposed structure has an improved aliasing rejection, and decreased passband droop in comparison with the corresponding comb filter. The aliasing rejection is increased by using a modified comb filters. Similarly, the passband droop is decreased by using a compensator at a low rate. The method is illustrated with one example. The structure has k stages, each decimated by 3, where k depends on decimation factor. Each stage is composed of the cascade of combs and modified combs. The structure based on polyphase decomposition is also presented. Some comparisons with methods from literature are also provided.

Thursday, June 20 14:30 - 15:50

Consumer Electronics for Motion Analysis 1go to top

Room: Sala Polveri
Chair: Francesco Di Nardo (Università Politecnica delle Marche, Italy)
Inertial Sensor Based Estimation of the Center of Pressure During Unperturbed Upright Stance
Alessandro Mengarelli, Stefano Cardarelli, Andrea Tigrini, Lorenzo Marchesini, Annachiara Strazza and Sandro Fioretti (Università Politecnica delle Marche, Italy); Federica Verdini (Universita' Politecnica delle Marche, Italy)
The unperturbed upright stance is considered one of the most important task for evaluating active and passive control of balance. In this field, the analysis of the center of pressure (COP) displacement is a valuable source of information for extracting and evaluating postural features in healthy and pathological subjects. COP is commonly measured by dynamometric force plates (DFP), which represent the gold-standard for measuring dynamic quantities. However, the availability of DFP is often limited by their cost and by the need for instrumented environments. In the past few years, inertial measurement units (IMU) have been proposed for estimating ground reaction forces and torques during selected motor tasks and under specific experimental conditions. Therefore, the aim of this study was to propose a novel method for deriving COP displacement during unperturbed posture by using a single sacrum-worn IMU. Results showed a remarkable correspondence between estimated COP time-series and those measured simultaneously by a DFP (root mean square error not beyond 1.5 mm, with a correlation not under 0.98). Further, COP the mean distance presented biases which are below 2 mm for both anteriorposterior and medial-lateral COP components, with the absence of any tendency towards proportional errors. Outcomes of this study seem to suggest the reliability of a single IMU based method for estimating COP displacement during balance maintenance task, providing a low-cost and unobtrusive procedure suitable for those contexts where classical motion analysis instrumentation is hardly or totally unavailable.
A 2D markerless gait analysis protocol to estimate the sagittal joint kinematics of children with cerebral palsy
Ilaria Cocchi (Sapienza University of Rome & University of Sassari, Italy); Giulio Figari (GPEM srl, Italy); Nicolò Valeri (University of Sassari, Italy); Gabriele Paolini (GPEM srl, Italy); Ugo Della Croce and Andrea Cereatti (University of Sassari, Italy); Jacques Riad (University of Gothenburg, Sweden); Evelina Pantzar (Örebro University Hospital, Sweden); Anders Magnuson (Örebro University, Sweden)
The quantitative analysis of human movement provides a deep understanding of the pathophysiological mechanisms underlying locomotion. The traditional marker-based stereo-photogrammetric systems and clinical protocols for motion analysis, although very accurate, have a number of disadvantages that limit their use to large clinical facilities. Among the disadvantages is the use of markers on the body, which can make the patient uneasy, especially children with cerebral palsy. To overcome the limitations of the marker-based stereophotogrammetry and to guarantee accuracy, reproducibility and usability of the measurement, a new markerless protocol is introduced, which, estimates the lower limb sagittal joint kinematics from RGB video images combined with measurements from an infrared depth (D) sensor. The validity of the markerless protocol is demonstrated by comparing the estimates obtained, with those resulting from the application of a common protocol applied to marker-based measurements. The interpretation of the differences found in this study should be treated carefully since they are the results not only of different measurement systems but also of different protocols (2D vs 3D).The proposed protocol for the estimation of the 2D joint kinematics of the lower limbs from RGB-D sensor data is a promising low-cost and simple solution for monitoring children with cerebral palsy.
A wearable device to assess postural sway
The maintenance of balance in upright stance is traditionally evaluated using heavy and expensive force platforms. The aim of this work is to prove the usefulness of a low-cost wearable sensor (an actigraph) to assess postural sway. We compared the performance of the device to a gold standard force platform. We analyzed measurements of postural sway in four conditions differently challenging the subject: with eyes open or closed, while keeping a small or large base of support. We estimated the main postural parameters (ellipse area, medio-lateral and antero-posterior root-mean square, eccentricity, sway path length) considering: 1) acceleration data recorded by the actigraph, and 2) traditional COP data obtained from the force platform. We found that it is possible to clearly distinguish the differences among the postural parameters, obtained in the various balance conditions, also using acceleration data. Our results show that the wearable device allows for obtaining information similar to those achievable by the force platform. This support the use of wearable devices to assess postural balance, in a handy and cheap manner.
Role of the Ground Reaction Force Components in Balance Assessment during Squatting Exercise performed by the Wii Balance Board
Alessandro Mengarelli, Andrea Tigrini, Lorenzo Marchesini, Annachiara Strazza, Sandro Fioretti and Stefano Cardarelli (Università Politecnica delle Marche, Italy); Federica Verdini (Universita' Politecnica delle Marche, Italy)
The Nintendo Wii Balance Board (WBB) has been proposed as a valid low-cost device for the assessment of balance in many different motor tasks. However, the direct comparisons with the gold-standard for balance measure represented by the laboratory grade force plates (FP) showed fixed biases affecting both the center of pressure (CoP) computation and its related spatial parameters, which can be partially attributed to the negligence of horizontal ground reaction force (HGRF) components in WBB-CoP computation. Thus, this study aimed to quantify to what extent the absence of HGRF affects the errors between FP and WBB in CoP related parameters during the squatting movement, where significant HGRFs are involved. Results showed that the HGRF negligence affects errors between FP and WBB to a significantly different extent when squat exercise is performed with two different techniques, being higher when squatting does not reach the maximal depth (average value of 60% versus 40% for all the considered parameters). Outcomes of this study seem to indicate that the use of WBB could be preferable for motor task with a marked mono-axial dynamics and in particular for the squatting exercise performed with a high squatting depth.

Thursday, June 20 16:10 - 16:40

Coffee Breakgo to top

Room: Break Area

Thursday, June 20 20:00 - 22:59

Gala Dinnergo to top

IL PASSETTO RESTAURANT, Piazza IV Novembre, https://www.ilpassettorestaurant.it/
Room: Outside

Friday, June 21

Friday, June 21 8:30 - 11:10

Registrationgo to top

Room: Registration Area

Friday, June 21 9:00 - 10:20

Consumer Healthcare 3go to top

Room: Sala Boxe
Chair: Alessandro Tonacci (Institute of Clinical Physiology (IFC) National Research Council of Italy (CNR), Italy)
Recurrence Quantification Analysis for Motion Artifacts in Wearable ECG Sensors
Amnah Nasim, Ilaria Marcantoni, Agnese Sbrollini and Micaela Morettini (Università Politecnica delle Marche, Italy); Laura Burattini (Polytechnic University of Marche, Italy)
Recurrence quantification analysis (RQA) allows the measurement of signal's regular and chaotic states using recurrence plots instead of deriving information purely from visual analysis. The current study presents RQA of multiple ECG time series simultaneously recorded through different electrodes and depicts the effect of motion artifacts through electrode synchronization and non-synchronization. The ECG data is acquired from a healthy 25-year-old male performing different exercise activities such as standing, walking and jumping. Also, the electrode in every recorded signal is placed at angle offset of 0, 45and 90. The RQA analysis measures recurrence rate (RR), line entropy (ENT) and average diagonal length (L) reveal a highly stable and least chaotic signal in case of standing (RR=0.73, ENT=4.94, L=106.12), somewhat stable and a bit chaotic in case of walking (RR=0.75, ENT=5.35, L=129.13) and least stable and most chaotic in case of subject performing a jump (RR=0.61, ENT=5.07, L=99.16). Secondly, highest and second highest disturbances with respect to exercise movements are observed for electrode combinations (3,4) and (1,4). Distinguishing values for RQA-based measures for different exercise movements suggest that RQA is a powerful tool for differentiation of regular and irregular states occurring due to motion artifacts in the temporal patterns of ECG.
Visible Light Communication by Using LED Array for Automatic Wheelchair Control in Hospital
Musashi Tsunoda and Chinthaka Premachandra (Shibaura Institute of Technology, Japan); H. A. Harindu Y. Sarathchandra, K. L. Amila Nuvan Perera and I. Tharindu Lakmal (Wayamba University of Sri Lanka, Sri Lanka); H. Waruna H. Premachandra (Wayamba University of Srilanka, Sri Lanka)
In Japan, there is a shortage of caretakers due to the increase of aging population. As a solution for this issue, an automated wheelchair driving system is proposed for the hospitals to reduce the workload of caretakers. In this system, the communication between the wheelchair and the external environment is realized by visible light communication which has no negative influence on the human body like other wave based communication methods. An LED matrix and a camera are used to achieve this communication. The wheelchair is guided by an LED blinking pattern which is recognized by a camera installed in the wheelchair. In this paper, the communication side of the system is studied. A set of experiments were carried out to validate the proposed system and their results emphasize the effectiveness of the system.
Monitoring patients with fragilities in the context of de-hospitalization services. An Ambient Assisted Living Healthcare Framework for e-Health applications
Emilio Sulis, Ilaria Angela Amantea, Guido Boella, Clara Cena, Roberta Fruttero and Sara Traina (University of Torino, Italy); Pierluigi De Cosmo and Lucrezia Armando (INFOLOGIC, Italy); Renata Marinello and Dario Bianca (City of Health and Science, Italy); Enrico Brunetti (University of Torino, Italy); Alessandra Bianco, Francesco Cattel and Mario Bo (City of Health and Science, Italy); Luca Carlo Feletti (CARETEK, Italy); Serena Ambrosini (CONSOFT, Italy)
This article focuses on e-health solutions for monitoring patients with fragilities in the context of de-hospitalization. Assistive technologies clearly have a direct and positive impact on the quality of life of patients, but they also improve the overall management of the organizational processes. In the first steps of "La Casa nel Parco" project funded by Regione Piemonte, we introduce here a general Ambient Assisted Living Healthcare framework to include business process analysis, telehealth and telemedicine applications. In particular, we define modeling and simulation of hospital services as a base to investigate the role of technological innovations. The project further investigates the impact of a telehealth solution with a wrist-worn device, as well as a telemedicine application based on augmented reality applied to pharmaceutical products. Our framework allows to explore the positive impact both on patient well-being as well as on business process management perspective.
MyDi application: towards automatic activity annotation of young patients with Type 1 diabetes
Lucia Migliorelli, Sara Moccia, Ismaela Avellino, Maria Chiara Fiorentino and Emanuele Frontoni (Università Politecnica delle Marche, Italy)
Type I diabetes mellitus (T1DM) is a widespread metabolic disorder characterized by pancreatic insufficiency. People with T1DM require: a lifelong insulin injection, to constantly monitor glycemia and to take note of their activities. This continuous follow-up, especially at a very young age, may be challenging. Adolescents with T1DM may develop anxiety symptoms and depression which can lead to the loss of glycemic control. An assistive technology that automatizes the activity monitoring process could support these young patients in managing T1DM. The aim of this work is to present the MyDi framework which integrates a smart glycemic diary (for Android users), to automatically record and store patient's activity via pictures and a deep-learning (DL)-based technology able to classify the activity performed by the patients (i.e., meal and sport) via picture analysis. The proposed approach was tested on two different datasets, the Insta-Dataset with 3498 pictures (also used for training and validating the DL model) and the MyDi- Dataset with 126 pictures, achieving very encouraging results in both cases (Preci= 1.0, Reci= 1.0, f1i= 1.0 with i in: [meal, sport]) prompting the possibility of translating this application in the T1DM monitoring process.

Consumer Electronics for Motion Analysis 2go to top

Room: Aula Didattica
Chair: Alessandro Mengarelli (Università Politecnica delle Marche, Italy)
Lifestyle analysis of a female group of university workers
Maintaining appropriate levels of Physical Activity (PA) and healthy lifestyles produces several health benefits. The U.S. Department of Health and Human Services recommended from 150 minutes to 300 minutes a week of moderate PA to obtain substantial health advantages. To promote well-being and healthy lifestyles, digital technologies play a very important role, since they can give a real-time feedback about the performed activities. In this study, we analyzed the lifestyle of a female group of university workers. We asked 23 healthy women belonging to the community of Politecnico di Torino to wear a device during a typical working day. The device was able to classify the activities performed by the subject in six classes: "resting" (i.e., sitting and laying), upright standing, walking, ascending stairs, descending stairs and "other activities" (comprising all the activities not included in the previous classes). We analyzed the time spent on each activity during the day and found that subjects spent, in average, almost one hour on dynamic activities (walking and stair climbing), that is in line with the recommendations. However, subjects did not carry out these activities continuously, but they split them into relatively short intervals whose maximum duration was approximately 10 minutes.
Evaluation of a New Wearable Technology for Kinematic Analysis During Perturbed Posture
Federica Verdini (Universita' Politecnica delle Marche, Italy); Alessandro Mengarelli, Stefano Cardarelli, Andrea Tigrini and Francesco Di Nardo (Università Politecnica delle Marche, Italy); Paola Di Florio (Università Politenica delle Marche, Italy); Gabriele Ligorio, Pietro Garofalo, Filippo Gianferrari and Michele Raggi (Turingsense, Italy); Sandro Fioretti (Università Politecnica delle Marche, Italy)
In this study a validity evaluation of a sensorized, full-body suit for kinematic measurement during perturbed upright stance task was proposed. Subjects underwent to a series of translational perturbations of the base of support in backward direction, with fixed speed and space (20 cm/s, 15 cm). The ankle joint angular displacement were simultaneously acquired by the full-body suit and an optoelectronic system, considered as the gold-standard for kinematics measurement. The two devices showed a good correspondence between measures, with absolute biases not higher than 0.8 degrees for both ankle joint range of motion and peaks amplitude. Results suggest the possibility to use the proposed wearable technology for assessing human body angular kinematics during perturbed posturography trials, in order to analyze balance strategies employed for external disruption withstanding and balance recovery.
People Movement Analysis with Automotive Radar
Ennio Gambi (Università Politecnica delle Marche, Italy); Adelmo De Santis (Università Politenica delle Marche, Italy); Gianluca Ciattaglia (Università Politecnica delle Marche, Italy)
Radars have become very important in the automotive field, as they can be used to increase safety and support autonomous driving. These devices work in the mmWave frequency range and allow a maximum range of 150m yet providing a great accuracy in distance measurement at a low cost. By processing the phase information taken from the radar is it possible to increase the precision of the radar and allow for a resolution in the order of tens of microns. In the present paper automotive radars are applied to the contactless monitoring of people walking, with the aim to evidence some peculiarities of the movement such as walking speed, length and speed of the steps, movement of legs and arms. The obtained results confirm the applicability of this radar in such a context, in order to obtain information to be used in machine learning application for an automatic identification of walking aspects.
Magnetometer-Free Sensor Fusion Applied to Pedestrian Tracking: A Feasibility Study
Stefano Cardarelli (Università Politecnica delle Marche, Italy); Paola Di Florio (Università Politenica delle Marche, Italy); Alessandro Mengarelli, Andrea Tigrini and Sandro Fioretti (Università Politecnica delle Marche, Italy); Federica Verdini (Universita' Politecnica delle Marche, Italy)
In this paper a feasibility study on magnetometer-free sensor-fusion in 2D pedestrian position and orientation tracking is presented. The sensor-fusion is performed through a nonlinear Kalman filter variant. The estimated orientation is then used to properly refer measured acceleration in a arbitrarily chosen navigation reference frame. Also, the same data is used to accurately remove the gravitational component from the measurement acceleration before any integration procedure to yield velocity and position. In addition, a custom curve for accurate foot-stance recognition is proposed. A velocity and acceleration reset is performed on the instants recognized thanks to the proposed time series. Estimated position of the foot-mounted IMU is then compared with fixed landmarks. Three walking tasks have been performed: a circle, a square and a longer indoor walk. Results show good accuracy in spite of the absence of heading information coming from the magnetometer (errors below 10 cm for the short trials), although on the long distances drift is more evident (2 m over 48 m of overall walking distance covered). Finally, a series of random walks have been performed to evaluate the foot stance recognition using the proposed time series against the simple accelerometer module evaluation.

Friday, June 21 10:20 - 11:10

Poster Sessiongo to top

Room: Poster Area
Chair: Francisco J. Bellido-Outeiriño (University of Córdoba, Spain)
Elders prefer female robots with a high degree of human likeness
Anna Esposito (Università della Campania Luigi Vanvitelli & and International Institute for Advanced Scientific Studies (IIASS), Italy); Terry Amorese (Università degli Studi della Campania Luigi Vanvitelli, Italy); Marialucia Cuciniello (Università degli Studi della Campania Luigi Vanvitell, Italy); Antonietta Esposito (Istituto Nazionale di Geofisica e Vulcanologia, Italy); Ilaria Pica, Maria Teresa Riviello, Alda Troncone and Gennaro Cordasco (Università degli Studi della Campania Luigi Vanvitelli, Italy)
Elders' acceptance of robots is still a novel field and a clear research methodology to assess users' preferences has not yet been developed. The exploitation of robots, as assistive technologies, requires the properly identification of users' needs and expectations and the matching of robot's role, appearance, and behavior to these needs. Robot's degree of resemblance to humans may play a fundamental role on their acceptance into domestic spheres. The present paper investigates elders' preferences towards female robots showing different levels of human likeness (two androids and a humanoid robot are involved in the study) considering their pragmatic, hedonic and attractive dimensions, as well as occupations elders entrusted to robots. A total of 51 elders (29 females) aged 65+ years were recruited. Participants were asked to watch video clips showing three speaking female manufactured robots (Erica, Sophia, and Pepper) and after each video clip, the Robot Appearance Questionnaire (RAQ) was administered. The results highlight that the degree of robot's human likeness affects elders' preferences in favor of android robots. Elders expressed a clear preference for female android rather than humanoid robots, in contrast with the current trend observed in literature. In addition, female robots were considered more suitable in performing housework rather than protection/security, healthcare, and front office occupations.
Evaluation of MCI Motor Performances During a Cognitive Dual Task Exercise
Gianmaria Mancioppi, Laura Fiorini and Maria Critelli (The Biorobotics Institute, Scuola Superiore Sant'Anna, Italy); Marco Timpano (USL Nordovest Toscana, Italy); Filippo Cavallo (The Biorobotics Institute, Scuola Superiore Sant'Anna, Italy); Erika Rovini (Scuola Superiore Sant'Anna, Italy)
The increased level of interdisciplinarity, in the field of neuropsychology, has led to the shift from the idea that walking represents an automatic and simple gesture, to the notion that it is a complex task involving several cognitive abilities. This appraisal occurred by the increased use of Cognitive and Motor Dual-Task paradigms, and by the growing interest in quantitative gait analysis. In this framework, we aim to investigate the relationship between cognition and motor skills, and to apply insights gathered from this study to the diagnostic process of Dementia and Mild Cognitive Impairment. Here it is proposed an innovative technological tool able to quantify gait parameters while providing for a Motor and Cognitive Dual-Task. The results of this research show a significant statistical difference between subjects suffering from Mild Cognitive Impairment with and without planning deficits.
A Cross-Protocol Proxy for Sensor Networks Based on CoAP
Smart technologies able to support people during their daily life, such as those linked to the Ambient Assisted Living (AAL) world, are gaining importance. In this regard, solutions to problems arising from the need to connect constrained networks with the internet are essential. In fact, constrained network, being a particular wireless network made of devices that have limited computational power and limited storage capacity, strongly differ from the internet network. Communication protocols try to overcome the issues related to the interconnection of smart devices in our smart world with the internet. A lot of efforts have been made in this direction, ending up with the creation of several different protocols and among them, in the applications context, the CoAP (Constrained Application Protocol) one is becoming increasingly relevant. The emergence of new protocols forces the need for developing proxy, a system able to intermediate between the two kind of networks and to translate between the relative protocols. In this paper we are going to present a cross-protocol proxy able to broker among the HTTP, MQTT and CoAP protocols and also able to implement the caching function, that as we will deepen, is an essential thing for the timeliness of communications. The proposed cross-protocol proxy has been tested under four operating conditions in terms of Throughput and Round Trip Time. The results show excellent performances for both the metrics taken into account, especially using the caching feature.
Postural stability evaluation using wearable wireless sensor
Paola Pierleoni, Alberto Belli, Lorenzo Palma, Michele Paoletti, Sara Raggiunto and Federica Pinti (Università Politecnica delle Marche, Italy)
This paper presents a wearable wireless sensor as a support tool used to assess the postural stability of a subject. Changes in postural steadiness are age-related. Their characterization and identification, through the development of a non-invasive device, will allow to identify persons at risk. In this work we develop an accurate device for real-time estimation of balance parameters with the aim to reduce falls in the elderly population. The proposed device embeds a 3-axis accelerometer, a 3-axis gyroscope and a 3-axis magnetometer. In order to get an accurate estimation of the displacement of the Center of Gravity (CoG) and of the Centre of Pressure (CoP), a data fusion algorithm is used. In particular, an Attitude and Heading Reference System (AHRS) has been implemented and optimized to provide an estimate of the orientation of the human body wearing the sensors. Moreover, we proposed an algorithm to evaluate the postural stability. Therefore, our algorithm has the aim of obtaining the CoG and CoP displacements in the anterior-posterior (AP) and medial-lateral (ML) directions, from the information provided by the AHRS, based on the inverse pendulum mathematical model. Results from these experiments on a population of healthy subjects suggest that the proposed device and algorithm can effectively be used for measuring CoG and CoP displacements.
Action recognition to estimate Activities of Daily Living (ADL) of elderly people
Matteo Gabrielli, Pietro Leo and Fabrizio Renzi (IBM Italia, Italy); Sonia Bergamaschi (University of Modena and Reggio Emilia, Italy)
This work proposes a method and preliminary experimental results to detect and recognize a set of Activities of Daily Living, carried out by elderly people in a residential context, by analyzing video of actions, recorded using a RGB Camera. The proposed solution is based on the creation of neural network models, in particular CNN, trained on data extracted and pre-processed from the "Moments in Time" dataset, a resource released by MIT-IBM Watson AI Lab that includes a collection of one million labeled 3 second videos from hundreds of categories.
Detecting User's Behavior Shift with Sensorized Shoes and Stigmergic Perceptrons
Paolo Barsocchi (ISTI-CNR, Italy); Nicola Carbonaro and Mario Giovanni C.A. Cimino (University of Pisa, Italy); Davide La Rosa (National Research Council, Italy); Filippo Palumbo (National Research Council (CNR), Italy); Alessandro Tognetti (Interdepartmental Research Center “Centro Piaggio”, University of Pisa, Italy); Gigliola Vaglini (University of Pisa, Italy)
As populations become increasingly aged, health monitoring has gained increasing importance. Recent advances in engineering of sensing, processing and artificial learning, make the development of non-invasive systems able to observe changes over time possible. In this context, the Ki-Foot project aims at developing a sensorized shoe and a machine learning architecture based on computational stigmergy to detect small variations in subjects gait and to learn and detect users behaviour shift. This paper outlines the challenges in the field and summarizes the proposed approach. The machine learning architecture has been developed and publicly released after early experimentation, in order to foster its application on real environments.
ErrP Signals Detection for Safe Navigation of a Smart Wheelchair
Lucio Ciabattoni (Universita' Politecnica delle Marche, Italy); Francesco Ferracuti and Alessandro Freddi (Università Politecnica delle Marche, Italy); Sabrina Iarlori (Universita' Politecnica delle Marche, Italy); Sauro Longhi and Andrea Monteriù (Università Politecnica delle Marche, Italy)
Assistive robots operate in complex environments and in presence of human beings, as such they are influenced by several factors which may lead to undesired outcomes: wrong sensor readings, unexpected environmental conditions or algorithmic errors represent just few examples. When the safety of the user must be guaranteed, a possible solution is to rely on a human-supervised approach. The proposed work presents a smart wheelchair, i.e. an electric powered wheelchair with semi-autonomous navigation capabilities, whose user is equipped with a Brain Computer Interface. During the wheelchair navigation, possible problems (e.g. obstacles) along the trajectory cause the generation of error-related potentials signals when noticed by the user. These signals are captured by the interface and are used to provide a feedback to the navigation task, in order to preserve safety and avoiding possible navigation issues.
Integrated Consumer Technologies for Older Adults' Quality of Life Improvement: the vINCI Project
Susanna Spinsante and Annachiara Strazza (Università Politecnica delle Marche, Italy); Ciprian Dobre (University Politehnica of Bucharest, Romania); Lidia Bajenaru (National Institute for Research and Development in Informatics, Romania); Constandinos X. Mavromoustakis (University of Nicosia & University of Nicosia Research Foundation, Cyprus); Jordi Mongay Batalla (Warsaw University of Technology & National Institute of Telecommunications, Poland); Piotr Krawiec (National Institute of Telecommunications & Warsaw University of Technology, Poland); George Georgescu (Connected Medical Devices, Romania); Gregor Molan (Comtrade Digital Services, Slovenia); Horacio Gonzalez-Velez (National College of Ireland, Ireland); Anna Marie Herghelegiu and Gabriel Ioan Prada (University of Medicine and Pharmacy Carol Davila, Romania); Rozeta Draghici (National Institute of Gerontology and Geriatrics Ana Aslan, Romania)
A low physical activity is one of the most important lifestyle risk factors for many chronic conditions in the older age. To tackle this issue, the consequences of which have relevant impact on the sustainability of national welfare in many developed countries, consumer technologies may be leveraged to offer easy-to-use tools enabling older adults to optimize their health-related quality of life, and promote an active and healthy longevity. This paper describes a technical platform named vINCI, obtained by the integration of consumer technologies with Assisted Living solutions and services, where multiple wearable devices work together to create an aggregated solution able to capture the various facets of events leading to the decrease in the perceived health-related quality of life, as typically associated with old age. Supporting the feedback of specialized medical evaluations, the vINCI technology enables elderly not only to self-evaluate their physical activity level, but also to change their behaviors and lifestyle in the long-term.
Electrocardiogram-Derived Respiratory Signal in Sleep Apnea by Segmented Beat Modulation Method
Agnese Sbrollini, Ilaria Marcantoni, Amnah Nasim and Micaela Morettini (Università Politecnica delle Marche, Italy); Laura Burattini (Polytechnic University of Marche, Italy)
The most common sleep disorder is sleep apnea, whose manifestations are long breathing pauses. Sleep apnea assessment is usually performed by polysomnography. During this long-term monitoring, patient respiration and other biosignals are recorded by many sensors, causing high level of discomfort. Thus, methods able to indirectly estimate the biosignal of interest from the others measured should be preferred. Respiration indirectly measured from electrocardiogram (ECG) is called ECG-derived respiratory (EDR) signal. Recently, Segmented Beat Modulation Method (SBMM) was proposed as a good method for EDR signal estimation in normal breathing. Thus, aim of this study was to assess the quality of EDR signal estimation by SBMM in pathological events of sleep apnea. With this purpose, sixteen long term polysomnographic recordings from MITBIH Polysomnographic Database were considered. After standard preprocessing, respiration and ECG signals were divided in 30s windows and, in order to match to provided annotations, each window was classified into Normal or Apnea. EDR signal was estimated by SBMM procedure from each ECG window. Respiration and EDR signals were then processed by Fourier analysis to extract respiration frequencies. Respiration frequencies computed from respiration and EDR signals were compared in term of error. Results confirmed the good quality of estimated EDR signal. Respiration frequency extracted from EDR signal in both Normal (16[13;19]cpm) and Apnea windows (18[15;21]cpm) are comparable to those extracted from respiration signal (Normal: 16 [13;19]cpm and Apnea: 18 [15;21]cpm), providing null error distributions. In conclusion, SBMM proved to be a promising tool for EDR signal estimation.
User-centered co-design and AGILE methodology for developing ambient assisting technologies
Francesco Tessarolo (University of Trento & Bruno Kessler Foundation, Italy); Giandomenico Nollo (University of Trento, Italy); Valentina Conotter (SocialIT srl, Italy); Giulia Onorati (Social IT srl, Italy); Evdokimos Konstantinidis and Despoina Petsani (Aristotle University of Thessaloniki, Italy); Panagiotis Bamidis (Medical School of the Aristotle University of Thessaloniki, Greece)
CAPTAIN project (Coach Assistant via Projected and TAngible INterface) is part of a research and innovation action, funded by the European Union, which aims at developing an innovative technology to help older adults overcoming frailties as they age. CAPTAIN will develop behavioral and Artificial Intelligence (AI) algorithms to provide personalized advice, guidance and follow-up to compensate for key age-related impairments during daily living, helping older adults to remain active and independent at their home. This will include risk avoidance, nutrition guidance, physical activity and cognitive training follow-up, guidance for lifestyle and social participation. A hybrid approach leveraging on concept from Design Thinking, Lean Startup approach and Scrum Agile framework will be followed by the project. The study plan, developed to coordinate and synchronize activities among all the different technological partners and living laboratories at pilot sites, is presented.
MRIndex: a tool for evaluating muscle involvement in neuromuscular diseases from MRI images
Daniela Marfisi and Maria Fantacci (University of Pisa, Italy); Raffaele Conte, Alessandro Tonacci and Francesco Sansone (Institute of Clinical Physiology (IFC) National Research Council of Italy (CNR), Italy); Guja Astrea and Filippo Santorelli (IRCCS Stella Maris Foundation, Italy)
The progress and severity of neuromuscular conditions can be monitored in several ways, most of which are invasive and, thus, poorly acceptable, for the patient. Among the least obtrusive solutions, the use of muscular magnetic resonance imaging is gaining importance in last decades, therefore becoming the elective methodology for defining the muscular involvement in such conditions. However, subjectivity is always quite frequent in the interpretation of biomedical images, and the diagnosis is often demanded to the experience of the clinician's eye. With MRIndex, a novel tool for the automated analysis of muscular magnetic resonance images, a quantitative "picture" of the muscular involvement in neuromuscular conditions is defined, stemming from the definition of fat infiltration within the muscular tissue, well-known biomarker for disease severity. The solution is currently used at clinics and preliminary at-a-glance results are quite promising. If such evidence is confirmed on larger cohorts and with a more robust statistical approach, the tool can represent a groundbreaking alternative in current clinical practice, possibly bringing to a more objective definition of the clinical status of a patient and helping in the personalization of the treatment, possibly improving their outcome.
Accuracy Evaluation of Force Measurement Through the Wii Balance Board During Squat and Functional Reach Tests
Federica Verdini (Universita' Politecnica delle Marche, Italy); Alessandro Mengarelli, Andrea Tigrini, Annachiara Strazza, Francesco Di Nardo, Stefano Cardarelli and Sandro Fioretti (Università Politecnica delle Marche, Italy)
The Nintendo Wii Balance Board (NWBB) is a commercial low cost device that can be used to measure the force exchanged with the ground and, in such a way, it constitute a valid alternative to expensive laboratory force plates (FP). Because of the lack of horizontal forces measure, the NWBB data resulted affected by inaccuracies when Center of Pressure (COP)-parameters are evaluated. The latter drawback affects in particular dynamic motor tasks, where greater forces are involved. Thus, this paper aimed to compare the NWBB and a FP to assess the error in the vertical force measure in two different highly dynamic tasks such as squatting (SQ) and functional reach test (FR). The findings confirmed the general agreement between the devices when used to measure force and related peaks during SQ and FR test. In particular it appears that the percentage errors are in line with those obtained for COP-parameters when evaluated without horizontal forces from FP, suggesting that this source of inaccuracy is device dependent and does not constitute a limitation of NWBB use in the case of single-device study.
Vulnerabilities and Attacks in a Smart Buildings Scenario
Rifat Seferi (University Politecnica delle Marche, Italy); Sofia Giangiacomi (COPRAT, Italy); Kejdi Berberi (Teleperformance, Albania)
Smart Buildings are an interconnected set of sensors, motors, controllers and different devices aiming to control the main functionalities of modern buildings. The evolution of these systems (from electronic and mechanical to complex elements) led to the exposition of smart buildings to new risks and threats. Moreover, the management of buildings by using wireless technologies (e.g, Bluetooth), present in the clear majority of the cases, is still complicated. In this paper, vulnerabilities and possible attacks to smart buildings are reported and described. This review could be a starting point to develop a method to simplify the management of buildings by using wireless technology.
A Web Application for Glasses Virtual Try-on in 3D Space
Davide Marelli (University of Milano-Bicocca, Italy); Simone Bianco (University of Milano Bicocca, Italy); Gianluigi Ciocca (University of Milano-Bicocca, Italy)
Virtual try-on is a technology that allow people to virtually check the appearance of accessories, makeup, hair style, hair color, clothes and potentially more on themselves. The virtual try-on presents many advantages over real try-on, it speeds up the process providing the possibility to test hundred of products without the need to go to the store. In this paper we propose a virtual try-on application specific for eyeglasses and sunglasses that can be easily used by the user by simply taking a picture of a face and selecting the desired frames. The try-on process is performed on a 3D face reconstructed from the input image allowing the user to see the virtual face and glasses from different view points. The try-on process is fully automated and does not require the user to provide anything else than the picture and selection of the glasses frames to test.

Coffee Breakgo to top

Room: Break Area

Friday, June 21 11:10 - 12:30

Automotivego to top

Room: Sala Boxe
Chair: Ivan Kastelan (University of Novi Sad, Faculty of Technical Sciences, Serbia)
Data Compression for Software Updating of ECUs
Naoki Suzuki (Kanagawa Institute of Technology, Japan); Toshiki Hayashi (Kanagwa Institute of Technology, Japan); Ryozo Kiyohara (Kanagawa Institute of Technology, Japan)
The number of electric control units (ECUs) in vehicles is increasing over the years along with an increase in the size of software. Because of complexity, it is difficult to release bug-free software in vehicles. Such software requires recalls that can be implemented via over-the-air (OTA) updating technology. OTA is important to implement this type of recall. The most important requirement of OTA updating is to minimize the software updating time, which can be applied with the binary difference technologies. However, in specific ECUs it is difficult to apply these technologies. In this study, we propose a new method of software updating for ECU that has a smaller size of RAM. This technology can reduce the time of software updating.
Overview of End-to-End Event Chain in Advanced Driver-Assistance Software following AUTOSAR
Marko Vulić (RT-RK Institute for Computer Based Systems, Bosnia and Herzegovina); Valentin Popović and Aleksandra Davidović (RT-RK Automotive, Serbia); Ivan Kastelan (University of Novi Sad, Faculty of Technical Sciences, Serbia)
In general, an event chain describes the casual order for a set of functionally dependent timing events. Each event chain has a well-defined stimulus and response, which describes its start and end point. In automotive software, that performs driver-assistance function, end-to-end event chain contains sensor acquisition as starting point, processing and decision in the middle, and action upon that decision in the end. Since performance and reliability of these functions have direct impact on road and driver safety, they have very strict safety requirements. To satisfy these requirements, system that performs these functions needs to have highly deterministic behavior and very low latency. End-to-End latency in this case means duration between the recognition of a sensor event and the corresponding action of an actuator. Correct description of end-to-end event chains enables accurate development of software architecture for specific driver-assistance functions and system in general. This paper presents an overview and general decomposition of event chains in automotive software developed according to AUTOSAR. It showcases major components of such event chain in layered AUTOSAR architecture, their respective time properties, and how it affects end-to-end latency.
Modeling and Development of AUTOSAR Software Components
Valentin Popović (RT-RK Automotive, Serbia); Marko Vulić (RT-RK Institute for Computer Based Systems, Bosnia and Herzegovina); Aleksandra Davidović (RT-RK Automotive, Serbia); Ivan Kastelan (University of Novi Sad, Faculty of Technical Sciences, Serbia)
In AUTOSAR architecture which has a component-oriented, hardware-independent software structure on application level, functionalities are implemented as individual units in a form of software components (SWCs). Because of their hardware independence, it is thus possible to develop SWCs without specific knowledge of the hardware used or planned, as well as to flexibly relocate existing SWCs to ECUs during development. This paper presents development of such SWCs from modelling to source code generation.
Optimize the Mild Hybrid Electric Vehicles control system to reduce the Emission
Marika Fanesi and David Scaradozzi (Università Politecnica delle Marche, Italy)
In the last decade virtual system integration and electrification have enabled an important change in automotive. The new technology and the attention about the environment, and renewable energy an increase in the use of electric power on road transports. The electrification process represents an im-portant growth, and the introduction of Mild Hybrid Electric Ve-hicles (MHEVs) is an effective solution to minimize and control the emission. In this paper, authors present the architecture used and the significant parameter considered. The preliminary results include the simulation of emission based on the New European Urban Cycle (NEDC) chosen as target testing cycle.

IoT 3go to top

Room: Aula Didattica
Chair: Chinthaka Premachandra (Shibaura Institute of Technology, Japan)
Detection of IoT Event Bursts in Smart Home Automation System
Sandra Ivanovic and Milica Matic (University of Novi Sad, Serbia); Marija Antic (University of Novi Sad); Istvan Papp (ObloLiving)
In this paper, one solution for detecting event bursts in smart home system is proposed. The goal is to identify user behavior patterns and eventually propose automation of frequently used actions, in order to improve user experience. Existing smart home system is extended with a data collection platform, which stores all device state changes in user's system. The stored data is then analyzed using Apache Spark, in order to detect event bursts, i.e. the changes that usually happen in a group. The details of the implementation are presented and the accuracy of the solution is tested on both artificial and real data test sets.
Performance analysis of indoor and outdoor real time localization system
Simone Orcioni, Massimo Conti, Francesco Gregorini and Pietro Antonelli (Università Politecnica delle Marche, Italy); Marco Giammarini, Rocco d'Aparo and Federico Robuffo (Filippetti S. r. l., Italy)
Many are the applications of indoor/outdoor localization sys-tems. This paper presents a localization device and algorithm based on UWB technology. The experimental measurements of an indoor and outdoor localization are presented.
Performance Evaluation of a Full IPv6-based Internet of Things Wireless Sensor Network
Lorenzo Incipini, Lorenzo Palma, Alberto Belli, Sara Raggiunto and Paola Pierleoni (Università Politecnica delle Marche, Italy)
In this paper we present an IPv6 Internet of Things Wireless Sensor Network whose complete architecture has been designed and developed by our research group. Every node of this Internet of Things solution is configured with its own global IPv6 address so making possible direct connections and queries to each node. A number of these nodes have been deployed with success in the Scrovegni Chapel heritage site (Padua, Italy) to monitor and control the lighting of the Giotto's frescoes. To analyze the performances of such IoT network we have previously developed an analog Wireless Sensor Network composed of the same type of nodes and deployed in a different indoor environment. We have used this network to measure the packet loss, round trip time and goodput over the IEEE 802.15.4 channel for 6LowPAN packets. To the best of our knowledge, this is the first operating Italian IPv6 global access Wireless Sensor Network.
An Evolutionary Approach for Line of Sight Relay Node Placement in A Sensor Network
MU Bigumjith Dias (University of Moratuwa, Sri Lanka); Gabriele Foresi (Universita' Politecnica delle Marche, Italy); Asitha L. Kulasekera and Katudampe Vithanage Damith Suresh Chathuranga (University of Moratuwa, Sri Lanka)
The wireless communication networks have been dominated the wired communication networks, throughout the past decades. In order to gain substantial performance most of the nodes have to be place in line of sight. Furthermore, line of sight communication is a must for visible light communication systems. Achieving the optimum relay node placement with line of sight communication configuration within unstructured real environment is quite a demanding goal. Authors have addressed the problem using an evolutionary approach in terms of maximizing the coverage and connectivity of the nodes. The novel algorithm is evaluated for 2D non-uniformly obstructed simulation environment.

Friday, June 21 12:35 - 13:15

Closing and Awardgo to top

Room: Sala Boxe

Friday, June 21 13:15 - 14:30

Lunchgo to top

Room: Break Area