1. Trang chủ
  2. » Kỹ Thuật - Công Nghệ

New Developments in Biomedical Engineering 2011 Part 18 potx

40 289 0
Tài liệu đã được kiểm tra trùng lặp

Đang tải... (xem toàn văn)

Tài liệu hạn chế xem trước, để xem đầy đủ mời bạn chọn Tải xuống

THÔNG TIN TÀI LIỆU

Thông tin cơ bản

Tiêu đề New Developments in Biomedical Engineering 2011 Part 18 potx
Tác giả Istrate, D., Vacher, M., Serignat, J.-F.
Chuyên ngành Biomedical Engineering
Thể loại Unknown
Năm xuất bản 2011
Định dạng
Số trang 40
Dung lượng 2,33 MB

Các công cụ chuyển đổi và chỉnh sửa cho tài liệu này

Nội dung

The integration of electronics and clothing is an emerging field which aims to the development of multi-functional, wearable electro-textiles for applications together with body function

Trang 2

Istrate, D., Vacher, M & Serignat, J.-F (2008) Embedded implementation of distress

situa-tion identificasitua-tion through sound analysis, The Journal on Informasitua-tion Technology in

Healthcare 6(3): 204–211.

Katz, S & Akpom, C (1976) A measure of primary sociobiological functions, International

Journal of Health Services 6(3): 493–508.

Kröse, B., van Kasteren, T., Gibson, C & van den Dool, T (2008) Care: Context awareness in

residences for elderly, Int Conference of the Int Soc for Gerontechnology, Pisa, Tuscany,

Italy

Kumiko, O., Mitsuhiro, M., Atsushi, E., Shohei, S & Reiko, T (2004) Input support for elderly

people using speech recognition, IEIC Technical Report 104(139): 1–6.

LeBellego, G., Noury, N., Virone, G., Mousseau, M & Demongeot, J (2006) A model for

the measurement of patient activity in a hospital suite, IEEE Trans on Information

Technology in Biomedicine 10(1): 92–99.

Litvak, D., Zigel, Y & Gannot, I (2008) Fall detection of elderly through floor vibrations and

sound, Proc 30th Annual Int Conference of the IEEE-EMBS 2008, pp 4632–4635.

Maunder, D., Ambikairajah, E., Epps, J & Celler, B (2008) Dual-microphone sounds of daily

life classification for telemonitoring in a noisy environment, Proc 30th Annual

Inter-national Conference of the IEEE-EMBS 2008, pp 4636–4639.

Michaut, F & Bellanger, M (2005) Filtrage adaptatif : théorie et algorithmes, Lavoisier.

Moore, D & Essa, I (2002) Recognizing multitasked activities from video using

stochas-tic context-free grammar, Proc of American Association of Artificial Intelligence (AAAI)

Conference 2002, Alberta, Canada.

Niessen, M., Van Maanen, L & Andringa, T (2008) Disambiguating sounds through context,

Proc Second IEEE International Conference on Semantic Computing, pp 88–95.

Noury, N., Hadidi, T., Laila, M., Fleury, A., Villemazet, C., Rialle, V & Franco, A (2008) Level

of activity, night and day alternation, and well being measured in a smart hospital

suite, Proc 30th Annual Int Conference of the IEEE-EMBS 2008, pp 3328–3331.

Noury, N., Villemazet, C., Barralon, P & Rumeau, P (2006) Ambient multi-perceptive

sys-tem for residential health monitoring based on electronic mailings experimentation

within the AILISA project, Proc 8th Int Conference on e-Health Networking, Applications

and Services HEALTHCOM 2006, pp 95–100.

Popescu, M., Li, Y., Skubic, M & Rantz, M (2008) An acoustic fall detector system that

uses sound height information to reduce the false alarm rate, Proc 30th Annual Int.

Conference of the IEEE-EMBS 2008, pp 4628–4631.

Portet, F., Fleury, A., Vacher, M & Noury, N (2009) Determining useful sensors for

auto-matic recognition of activities of daily living in health smart home, in Intelligent Data

Analysis in Medicine and Pharmacology (IDAMAP2009), Verona, Italy

Rabiner, L & Luang, B (1996) Digital processing of speech signals, Prentice-Hall.

Renouard, S., Charbit, M & Chollet, G (2003) Vocal interface with a speech memory for

dependent people, Independent Living for Persons with Disabilities pp 15–21.

Rialle, V., Ollivet, C., Guigui, C & Hervé, C (2008) What do family caregivers of alzheimer’s

disease patients desire in health smart home technologies? contrasted results of a

wide survey, Methods of Information in Medicine 47: 63–69.

Saeys, Y., Inza, I & Larrañaga, P (2007) A review of feature selection techniques in

bioinfor-matics, Bioinformatics 23: 2507–2517.

Soo, J.-S & Pang, K (1990) Multidelay block frequency domain adaptive filter, IEEE Trans on

Acoustics, Speech and Signal Processing 38(2): 373–376.

Takahashi, S.-y., Morimoto, T., Maeda, S & Tsuruta, N (2003) Dialogue experiment for elderly

people in home health care system, Text Speech and Dialogue (TSD) 2003.

Tran, Q T & Mynatt, E D (2003) What was i cooking? towards déjà vu displays of everyday

memory, Technical report.

Vacher, M., Fleury, A., Serignat, J.-F., Noury, N & Glasson, H (2008) Preliminary

evalua-tion of speech/sound recognievalua-tion for telemedicine applicaevalua-tion in a real environment,

The 9thAnnual Conference of the International Speech Communication Association, SPEECH’08 Proceedings, Brisbane, Australia, pp 496–499.

INTER-Vacher, M., Serignat, J.-F & Chaillol, S (2007) Sound classification in a smart room

environ-ment: an approach using GMM and HMM methods, Advances in Spoken Language Technology, SPED 2007 Proceedings, Iasi, Romania, pp 135–146.

Vacher, M., Serignat, J.-F., Chaillol, S., Istrate, D & Popescu, V (2006) Speech and sound

use in a remote monitoring system for health care, Lecture Notes in Computer Science, Artificial Intelligence, Text Speech and Dialogue, vol 4188/2006, Brno, Czech Republic,

pp 711–718

Valin, J.-M (2007) On adjusting the learning rate in frequency domain echo cancellation with

double talk, IEEE Trans on Acoustics, Speech and Signal Processing 15(3): 1030–1034.

Valin, J.-M & Collings, I B (2007) A new robust frequency domain echo canceller with

closed-loop learning rate adaptation, IEEE Int Conference on Acoustics, Speech and nal Processing, ICASSP’07 Proceedings Vol 1, Honolulu, Hawaii, USA, pp 93–96 Vaseghi, S V (1996) Advanced Signal Processing and Digital Noise Reduction, 1996.

Sig-Vaufreydaz, D., Bergamini, C., Serignat, J.-F., Besacier, L & Akbar, M (2000) A new

method-ology for speech corpora definition from internet documents, LREC’2000, 2nd Int Conference on Language Ressources and Evaluation, Athens, Greece, pp 423–426.

Wang, J.-C., Lee, H.-P., Wang, J.-F & Lin, C.-B (2008) Robust environmental sound

recogni-tion for home automarecogni-tion, IEEE Trans on Automarecogni-tion Science and Engineering 5(1): 25–

31

Wilpon, J & Jacobsen, C (1996) A study of speech recognition for children and the elderly,

IEEE Int Conference on Acoustics, Speech and Signal Processing, pp 349–352.

Trang 3

Complete Sound and Speech Recognition System for Health Smart Homes: Application to the Recognition of Activities of Daily Living 673

Istrate, D., Vacher, M & Serignat, J.-F (2008) Embedded implementation of distress

situa-tion identificasitua-tion through sound analysis, The Journal on Informasitua-tion Technology in

Healthcare 6(3): 204–211.

Katz, S & Akpom, C (1976) A measure of primary sociobiological functions, International

Journal of Health Services 6(3): 493–508.

Kröse, B., van Kasteren, T., Gibson, C & van den Dool, T (2008) Care: Context awareness in

residences for elderly, Int Conference of the Int Soc for Gerontechnology, Pisa, Tuscany,

Italy

Kumiko, O., Mitsuhiro, M., Atsushi, E., Shohei, S & Reiko, T (2004) Input support for elderly

people using speech recognition, IEIC Technical Report 104(139): 1–6.

LeBellego, G., Noury, N., Virone, G., Mousseau, M & Demongeot, J (2006) A model for

the measurement of patient activity in a hospital suite, IEEE Trans on Information

Technology in Biomedicine 10(1): 92–99.

Litvak, D., Zigel, Y & Gannot, I (2008) Fall detection of elderly through floor vibrations and

sound, Proc 30th Annual Int Conference of the IEEE-EMBS 2008, pp 4632–4635.

Maunder, D., Ambikairajah, E., Epps, J & Celler, B (2008) Dual-microphone sounds of daily

life classification for telemonitoring in a noisy environment, Proc 30th Annual

Inter-national Conference of the IEEE-EMBS 2008, pp 4636–4639.

Michaut, F & Bellanger, M (2005) Filtrage adaptatif : théorie et algorithmes, Lavoisier.

Moore, D & Essa, I (2002) Recognizing multitasked activities from video using

stochas-tic context-free grammar, Proc of American Association of Artificial Intelligence (AAAI)

Conference 2002, Alberta, Canada.

Niessen, M., Van Maanen, L & Andringa, T (2008) Disambiguating sounds through context,

Proc Second IEEE International Conference on Semantic Computing, pp 88–95.

Noury, N., Hadidi, T., Laila, M., Fleury, A., Villemazet, C., Rialle, V & Franco, A (2008) Level

of activity, night and day alternation, and well being measured in a smart hospital

suite, Proc 30th Annual Int Conference of the IEEE-EMBS 2008, pp 3328–3331.

Noury, N., Villemazet, C., Barralon, P & Rumeau, P (2006) Ambient multi-perceptive

sys-tem for residential health monitoring based on electronic mailings experimentation

within the AILISA project, Proc 8th Int Conference on e-Health Networking, Applications

and Services HEALTHCOM 2006, pp 95–100.

Popescu, M., Li, Y., Skubic, M & Rantz, M (2008) An acoustic fall detector system that

uses sound height information to reduce the false alarm rate, Proc 30th Annual Int.

Conference of the IEEE-EMBS 2008, pp 4628–4631.

Portet, F., Fleury, A., Vacher, M & Noury, N (2009) Determining useful sensors for

auto-matic recognition of activities of daily living in health smart home, in Intelligent Data

Analysis in Medicine and Pharmacology (IDAMAP2009), Verona, Italy

Rabiner, L & Luang, B (1996) Digital processing of speech signals, Prentice-Hall.

Renouard, S., Charbit, M & Chollet, G (2003) Vocal interface with a speech memory for

dependent people, Independent Living for Persons with Disabilities pp 15–21.

Rialle, V., Ollivet, C., Guigui, C & Hervé, C (2008) What do family caregivers of alzheimer’s

disease patients desire in health smart home technologies? contrasted results of a

wide survey, Methods of Information in Medicine 47: 63–69.

Saeys, Y., Inza, I & Larrañaga, P (2007) A review of feature selection techniques in

bioinfor-matics, Bioinformatics 23: 2507–2517.

Soo, J.-S & Pang, K (1990) Multidelay block frequency domain adaptive filter, IEEE Trans on

Acoustics, Speech and Signal Processing 38(2): 373–376.

Takahashi, S.-y., Morimoto, T., Maeda, S & Tsuruta, N (2003) Dialogue experiment for elderly

people in home health care system, Text Speech and Dialogue (TSD) 2003.

Tran, Q T & Mynatt, E D (2003) What was i cooking? towards déjà vu displays of everyday

memory, Technical report.

Vacher, M., Fleury, A., Serignat, J.-F., Noury, N & Glasson, H (2008) Preliminary

evalua-tion of speech/sound recognievalua-tion for telemedicine applicaevalua-tion in a real environment,

The 9thAnnual Conference of the International Speech Communication Association, SPEECH’08 Proceedings, Brisbane, Australia, pp 496–499.

INTER-Vacher, M., Serignat, J.-F & Chaillol, S (2007) Sound classification in a smart room

environ-ment: an approach using GMM and HMM methods, Advances in Spoken Language Technology, SPED 2007 Proceedings, Iasi, Romania, pp 135–146.

Vacher, M., Serignat, J.-F., Chaillol, S., Istrate, D & Popescu, V (2006) Speech and sound

use in a remote monitoring system for health care, Lecture Notes in Computer Science, Artificial Intelligence, Text Speech and Dialogue, vol 4188/2006, Brno, Czech Republic,

pp 711–718

Valin, J.-M (2007) On adjusting the learning rate in frequency domain echo cancellation with

double talk, IEEE Trans on Acoustics, Speech and Signal Processing 15(3): 1030–1034.

Valin, J.-M & Collings, I B (2007) A new robust frequency domain echo canceller with

closed-loop learning rate adaptation, IEEE Int Conference on Acoustics, Speech and nal Processing, ICASSP’07 Proceedings Vol 1, Honolulu, Hawaii, USA, pp 93–96 Vaseghi, S V (1996) Advanced Signal Processing and Digital Noise Reduction, 1996.

Sig-Vaufreydaz, D., Bergamini, C., Serignat, J.-F., Besacier, L & Akbar, M (2000) A new

method-ology for speech corpora definition from internet documents, LREC’2000, 2nd Int Conference on Language Ressources and Evaluation, Athens, Greece, pp 423–426.

Wang, J.-C., Lee, H.-P., Wang, J.-F & Lin, C.-B (2008) Robust environmental sound

recogni-tion for home automarecogni-tion, IEEE Trans on Automarecogni-tion Science and Engineering 5(1): 25–

31

Wilpon, J & Jacobsen, C (1996) A study of speech recognition for children and the elderly,

IEEE Int Conference on Acoustics, Speech and Signal Processing, pp 349–352.

Trang 5

New emerging biomedical technologies for home-care and telemedicine applications: the Sensorwear project

Luca Piccini, Oriana Ciani and Giuseppe Andreoni

X

New emerging biomedical technologies for

home-care and telemedicine applications:

the Sensorwear project

Luca Piccini, Oriana Ciani and Giuseppe Andreoni

Politecnico di Milano, INDACO Department

Italy

1 Introduction

The Grey Booming phenomenon is one of the major issues indicated by the European Union

as a problem to be analysed and faced by the Seventh Framework Programme (FP7)

Statistics highlighted that elderly people (over 65 years old) should double in the next 40

years The medical and health care to such an ‘older’ society means growing expenditures

for the UE national health systems, which already amount to significant percentages of the

Gross Domestic Product (GDP) in the different countries The UE Healthcare Systems risk to

collapse if strong countermeasures will not be undertaken Agreeing with this assumption,

the European Commission included among its priorities the stimuli to deeply remodel the

national healthcare systems France, United Kingdom, Holland, Austria, Italy and other

countries drafted national programs in order to face this emerging problem More in detail,

cardiac and respiratory diseases have been identified as some of the most frequent causes of

hospitalization; telemedicine and home-care have been therefore selected to face the

negative evolution of these pathologies, both in clinical and economical terms, assuring

domestic assistance for older people as well as disabled or chronic patients The rationale of

this choice is the opportunity of reducing the overall costs while maintaining high quality of

care and providing an easy access to care from any place, at any time Moreover the focus of

healthcare consequently shifts from treatment to prevention and early diagnosis, thanks to

the contribution of parallel wellness programs, too

Increasing the impact of home-care solutions is a difficult challenge, since technological

issues, such as biosignals monitoring, data communications and basic automated signal

analysis coexist with the efforts to improve new technologies’ acceptability by the patients,

who need to interact with them for long time Generally these users are not technologically

skilled therefore textile sensors platforms represent an ideal way to develop the

telemedicine approach

Under these perspectives, the research and development of Wearable Health Systems

(WHS) become even relevant They are expected to play a significant role on the spreading

of ‘extra-hospital’ cares, thus improving the national health policies effectiveness and the

citizens’ quality of life, too

34

Trang 6

WHS are integrated systems on body-worn platforms, such as wrist-worn devices or biomedical clothes, offering pervasive solutions for continuous health status monitoring trough non-invasive biomedical, biochemical and physical measurements (Lymberis & Gatzoulis, 2006) In other words, they provide not only a remote monitoring platform for prevention and early diagnosis, but also a valid contribution to disease management and support of elderly or people in need; in particular, they enable multi-parametric monitoring including body-kinematics, vital signs, biochemical as well as emotional and sensorial parameters in a defined social and environmental context

The integration of electronics and clothing is an emerging field which aims to the development of multi-functional, wearable electro-textiles for applications together with body functions monitoring, actuation, communication, data transfer and individual environment control Furthermore, the integration of advanced microsystems at the fibre core, in conjunction with user interfaces, power sources and embedded software, make R&D

in this field extremely challenging Moreover, current research is dealing with the development of stretchable conductive patterns and soft-touch substrates for component textile mounting and interconnection

As a matter of fact, WHS cope with a variety of challenging topics, whose complexity increases with their integration: wireless communication, power supply and management, data processing, new algorithms for biosignal analysis, connection, sensors’ cleaning and stability over time and external conditions, sensors positioning on the human body, user’s interface, garment’s elasticity and adherence to the skin and other minor themes

Surely the first issue to be managed is the technological one - current state of the art has achieved a good level of maturity to be industrialized and brought to the market - but another key factor, that is still not mature enough, is the ergonomic or human factor in terms

of device’s usability, comfort and acceptance by the end user According to the authors,

design for wearability is necessary for the real and definitive acknowledgment of WHS in

clinical applications, telemedicine and more (Andreoni, 2008) That’s the reason why, besides the main objectives of developing healthcare wearable devices, meeting the aforementioned requirements for enhanced user-friendliness, affordability and unobtrusive monitoring in several clinical applications is becoming a growing topic of worldwide research about WHS

In order to let WHS regularly break into the healthcare practice this and other issues should

be solved, for example, from the commercial and industrial point of view, the consolidation

of R&D results in different domains and their integration (David, 2007) The Sensorwear

project tries to organically coordinate the emerging technologies in the field of wearable biomedical devices, conductive yarns or garments, embedded monitoring devices, automated alarm systems and ICT channels optimizations, in order to design a complete, automated service for home and clinical cardiac monitoring applications

2 The international scenario of wearable telemonitoring systems

Wearable solutions for biophysical conditions monitoring can address many of the emerging issues previously described for a broad cross-section of user groups Elderly care and disease management are just the immediate application, in addition to wellness and sport which represent significant segments that can benefit from continuous, remote and personal monitoring solutions

Trang 7

WHS are integrated systems on body-worn platforms, such as wrist-worn devices or

biomedical clothes, offering pervasive solutions for continuous health status monitoring

trough non-invasive biomedical, biochemical and physical measurements (Lymberis &

Gatzoulis, 2006) In other words, they provide not only a remote monitoring platform for

prevention and early diagnosis, but also a valid contribution to disease management and

support of elderly or people in need; in particular, they enable multi-parametric monitoring

including body-kinematics, vital signs, biochemical as well as emotional and sensorial

parameters in a defined social and environmental context

The integration of electronics and clothing is an emerging field which aims to the

development of multi-functional, wearable electro-textiles for applications together with

body functions monitoring, actuation, communication, data transfer and individual

environment control Furthermore, the integration of advanced microsystems at the fibre

core, in conjunction with user interfaces, power sources and embedded software, make R&D

in this field extremely challenging Moreover, current research is dealing with the

development of stretchable conductive patterns and soft-touch substrates for component

textile mounting and interconnection

As a matter of fact, WHS cope with a variety of challenging topics, whose complexity

increases with their integration: wireless communication, power supply and management,

data processing, new algorithms for biosignal analysis, connection, sensors’ cleaning and

stability over time and external conditions, sensors positioning on the human body, user’s

interface, garment’s elasticity and adherence to the skin and other minor themes

Surely the first issue to be managed is the technological one - current state of the art has

achieved a good level of maturity to be industrialized and brought to the market - but

another key factor, that is still not mature enough, is the ergonomic or human factor in terms

of device’s usability, comfort and acceptance by the end user According to the authors,

design for wearability is necessary for the real and definitive acknowledgment of WHS in

clinical applications, telemedicine and more (Andreoni, 2008) That’s the reason why,

besides the main objectives of developing healthcare wearable devices, meeting the

aforementioned requirements for enhanced user-friendliness, affordability and unobtrusive

monitoring in several clinical applications is becoming a growing topic of worldwide

research about WHS

In order to let WHS regularly break into the healthcare practice this and other issues should

be solved, for example, from the commercial and industrial point of view, the consolidation

of R&D results in different domains and their integration (David, 2007) The Sensorwear

project tries to organically coordinate the emerging technologies in the field of wearable

biomedical devices, conductive yarns or garments, embedded monitoring devices,

automated alarm systems and ICT channels optimizations, in order to design a complete,

automated service for home and clinical cardiac monitoring applications

2 The international scenario of wearable telemonitoring systems

Wearable solutions for biophysical conditions monitoring can address many of the

emerging issues previously described for a broad cross-section of user groups Elderly care

and disease management are just the immediate application, in addition to wellness and

sport which represent significant segments that can benefit from continuous, remote and

personal monitoring solutions

During the last years, different research projects all over the European Union were dedicated to the creation of telemonitoring systems based on wearable or standard sensors MyHeart is one of the most important and complete among them Notwithstanding the relevant efforts that have been made since 2000 by the granted projects of the Seventh Framework Program (FP7), researchers and industries are still trying to improve patients’ condition monitoring at home using unobtrusive sensors built into everyday objects able to automatically report to clinicians1 (Lymberys & De Rossi, 2004) These examples and other projects demonstrated both the importance of such applications and the technological problems related to the creation of such a systems

On the other side, pilot studies were lead in order to evaluate the potential impact of care monitoring in terms of costs through a comparison with standard instrumentation The

home-EU Commission, in fact, has underlined the economical potentialities of such solutions, but also has pointed out doubts with the achievement of the potential results and the effective introduction of these technologies in the healthcare systems (COM 689, 2008)

All the predictive models, analysis and studies confirmed the importance of the wearable telemonitoring scenario, but many problems occur if one aims to the implementation of an industrial project and not only to a research prototype (Lymberis & Paradiso, 2008)

The Sensorwear project tries to avoid the segmentation of technologies and competences, concentrating a small, skilled group of people for the creation of a wearable, unobtrusive, low cost and fully automated solution whose usability, reliability and release of brief information are the most peculiar qualities

The market analysis has shown there are no commercial solutions able to assure those requirements with a complete wearable system for daily clinical monitoring It is not uncommon reading about prototypes or finding patents about wearable systems for health care and catching poor information coming from military applications context, not accessible by definition (Pantepopulous & Bourbakis, 2008) To date, the main companies involved in the development of wearable monitoring systems such as Body Media Inc., Sensatex Inc., Textronics Inc and Vivometrics Inc., experience every day the need for more consistent and remote monitoring of individuals for a variety of purpose: from elderly care

to chronic disease management and others Their solutions are just beginning the transition from the development phase into commercialization, facing the barrier of the regulatory approval, which remains critical for many of the producers

Just to give an example, a common electrocardiograph, the instrument allowing the execution of an electrocardiogram exam, costs about 600€ in the UE market and cannot be used with wearable sensors to provide unobtrusive measures The paradigm of measures transparency requires solutions’ refinement or improvement or the design of new integrated systems when noise, artefacts or ergonomic deficiencies enlarge

The Sensorwear system points at achieving these crucial objectives

1 For more details, go to: http://heartcycle.med.auth.gr and http://www.ehealthnews.eu

Trang 8

3 The Sensorwear project

The Sensorwear project focuses on design and development of a low-cost, industrial solution for smart home-monitoring and hospital applications The objective is not to create

a life-support system, but a reliable, cost effective solution able to monitor biosignals detecting specific conditions requested by clinicians and to transmit them consequently through a long-range communication channel

The project is granted by the Regione Lombardia and it involves the Politecnico di Milano - INDACO Department -, three technological partners (STMicroelectronics, Microsystems and SXT-Sistemi per Telemedicina), one clothes manufacturer (MCS – Manifatture Cotoniere Settentrionali), a service provider company for the textile and clothing sector (Centro Tessile Cotoniero) and the Mater Domini Hospital in Castellanza (IT), the project’s clinical partner

We will illustrate the main aspects related to the project’s objectives, technical solutions, applications and expected results in the following paragraphs

3.1 Objectives and overall architecture

The Sensorwear project aims at developing a complete home monitoring service able to collect a set of different biosignals in a transparent way during the spontaneous activity of the subjects: this paradigm is known as unobtrusive measure An important part of the project is the creation of a Body Sensor Network (BSN) dedicated to the health state monitoring trough record, process and transmission of the biosignals and some useful parameters obtained from them BSN is mainly based on wearable sensors for the collection

of biopotentials (like the electrocardiographic signals, the ECG) and integrated and miniaturized electronic solution based on Bluetooth® technology

The detailed objectives of the project are (Fig 1):

 Research, development and production of a System in Package (SIP) solution for monitoring, processing and transmission

 Research, development and production of embedded sensors

 Creation of fully featured t-shirts with integrated SIP devices to be tested and used both at home and in hospital

 Development of software and algorithms for the processing and management of signals, data and alarm for the different applications

 Development of software for remote data receiving and database integration

In order to fulfil those items, the fundamental point of the product industrialisation, which

is a peculiarity of the Sensorwear project, is continuously kept into consideration In this way, the final solution is expected to be compliant with the specifications for medical devices of class IIa Garments’ testing, which is an ongoing concern, is an unavoidable step

in order to ensure biocompatibility

The architecture of the system is essentially composed by four main systems:

1 t-shirt with embedded electrodes for the collection of bio-potentials

2 preconditioning and acquisition system

3 processing and transmission device

4 remote data management software

The second and third systems compose a body gateway able to directly control a mobile phone without requiring the user interaction Actually, the possibility to act in a fully automated way is another significant feature of the Sensorwear device

Trang 9

3 The Sensorwear project

The Sensorwear project focuses on design and development of a low-cost, industrial

solution for smart home-monitoring and hospital applications The objective is not to create

a life-support system, but a reliable, cost effective solution able to monitor biosignals

detecting specific conditions requested by clinicians and to transmit them consequently

through a long-range communication channel

The project is granted by the Regione Lombardia and it involves the Politecnico di Milano -

INDACO Department -, three technological partners (STMicroelectronics, Microsystems and

SXT-Sistemi per Telemedicina), one clothes manufacturer (MCS – Manifatture Cotoniere

Settentrionali), a service provider company for the textile and clothing sector (Centro Tessile

Cotoniero) and the Mater Domini Hospital in Castellanza (IT), the project’s clinical partner

We will illustrate the main aspects related to the project’s objectives, technical solutions,

applications and expected results in the following paragraphs

3.1 Objectives and overall architecture

The Sensorwear project aims at developing a complete home monitoring service able to

collect a set of different biosignals in a transparent way during the spontaneous activity of

the subjects: this paradigm is known as unobtrusive measure An important part of the

project is the creation of a Body Sensor Network (BSN) dedicated to the health state

monitoring trough record, process and transmission of the biosignals and some useful

parameters obtained from them BSN is mainly based on wearable sensors for the collection

of biopotentials (like the electrocardiographic signals, the ECG) and integrated and

miniaturized electronic solution based on Bluetooth® technology

The detailed objectives of the project are (Fig 1):

 Research, development and production of a System in Package (SIP) solution for

monitoring, processing and transmission

 Research, development and production of embedded sensors

 Creation of fully featured t-shirts with integrated SIP devices to be tested and used

both at home and in hospital

 Development of software and algorithms for the processing and management of

signals, data and alarm for the different applications

 Development of software for remote data receiving and database integration

In order to fulfil those items, the fundamental point of the product industrialisation, which

is a peculiarity of the Sensorwear project, is continuously kept into consideration In this

way, the final solution is expected to be compliant with the specifications for medical

devices of class IIa Garments’ testing, which is an ongoing concern, is an unavoidable step

in order to ensure biocompatibility

The architecture of the system is essentially composed by four main systems:

1 t-shirt with embedded electrodes for the collection of bio-potentials

2 preconditioning and acquisition system

3 processing and transmission device

4 remote data management software

The second and third systems compose a body gateway able to directly control a mobile

phone without requiring the user interaction Actually, the possibility to act in a fully

automated way is another significant feature of the Sensorwear device

Fig 1 Sensorwear main activities: different tasks and their relationships

The signals identified for the specific purpose of the telecardiology application are:

 Three ECG leads

 Body movement

 Respiratory frequency

 Cardiac output monitoring

The ECG signal is the most important one allowing the device to detect useful parameters like Heart Rate (HR), arrhythmias and their classification, ST line anomalies ECG is a primary source of indications about health condition, so it receives, at least at early stages, greater attention

The body movement is recorded through a three-axial accelerometer, whose properly processed signals allow determining the number of steps and the body position with respect

to the earth gravity

Furthermore, it is possible to detect the changing of the cardiac output during the day and also the respiratory movements through impedance cardiography measure (ICG)

3.2 Technological key-point and main issues

The main objectives of this project deal with the creation of an industrial, compact, easy to use, automated solution designed with a special attention to elderly people These

INDUSTRIAL SPECIFICATIONS

& FEASIBILITY

MATERIALS &

METHODS for the WEARABLE COMPONENTS PRODUCTION

SIP DESIGN &

PROTOTYPING SERVICES & SYSTEM

Clinical Evaluation

R&D AND INDUSTRIALIZATION

VALIDATION

Trang 10

demanding requirements are addressed by the different skills of the partners on yarns and textile solutions, electronic design and production, data collection and databases

WEARABLE SENSORS

The t-shirts were designed in order to facilitate the integration of sensors during the industrialization and to assure the best sensors’ positioning for ECG and ICG signals quality The design of garments is a crucial point in the field of unobtrusive measures, in fact as previous research projects and studies have evidenced, it needs configurations able to reduce the effects of movements, without impacting the comfort The testing phase for the t-shirt, their sensors and sensors’ position has started with the ECG signal check, following a specific protocol First of all the signals are recorded with the first prototype and standard electrode in the Einthoven's configuration, afterwards the device has to collect signals through the t-shirt The last scheduled test requires to connect the prototype to standard electrodes but placed in the same positions of the wearable ones Each recording is done 3 minutes at rest and 3 minutes into action

ERGONOMIC and MECHANICAL ASPECTS

As far as the design of the t-shirts and the adherence of sensors affect the quality of the ECG and ICG signals, the enclosure of the device and its connection to the sensors pathways strongly impact on both the usability and the industrial sustainability Our analysis of the production process and its related constraints evidenced the necessity to conceive custom boxes in order to create a real comfortable solution without renouncing to an appealing product The enclosure will also include the visual signalling with yellow and green leds, compliant to the specifications for Holter medical devices (Fig 2)

Moreover the custom case can be inserted in a docking station, directly sewed to the t-shirt and including the sensors connectors

At this purpose, a custom solution is not a cost-effective one, while the use of the docking station as mating support can allow the choice of stable, reliable although simple and cheaper connectors

Fig 2 The ergonomic study for the user interface and the device shape as regards wearability issues

Trang 11

demanding requirements are addressed by the different skills of the partners on yarns and

textile solutions, electronic design and production, data collection and databases

WEARABLE SENSORS

The t-shirts were designed in order to facilitate the integration of sensors during the

industrialization and to assure the best sensors’ positioning for ECG and ICG signals

quality The design of garments is a crucial point in the field of unobtrusive measures, in

fact as previous research projects and studies have evidenced, it needs configurations able to

reduce the effects of movements, without impacting the comfort The testing phase for the

t-shirt, their sensors and sensors’ position has started with the ECG signal check, following a

specific protocol First of all the signals are recorded with the first prototype and standard

electrode in the Einthoven's configuration, afterwards the device has to collect signals

through the t-shirt The last scheduled test requires to connect the prototype to standard

electrodes but placed in the same positions of the wearable ones Each recording is done 3

minutes at rest and 3 minutes into action

ERGONOMIC and MECHANICAL ASPECTS

As far as the design of the t-shirts and the adherence of sensors affect the quality of the ECG

and ICG signals, the enclosure of the device and its connection to the sensors pathways

strongly impact on both the usability and the industrial sustainability Our analysis of the

production process and its related constraints evidenced the necessity to conceive custom

boxes in order to create a real comfortable solution without renouncing to an appealing

product The enclosure will also include the visual signalling with yellow and green leds,

compliant to the specifications for Holter medical devices (Fig 2)

Moreover the custom case can be inserted in a docking station, directly sewed to the t-shirt

and including the sensors connectors

At this purpose, a custom solution is not a cost-effective one, while the use of the docking

station as mating support can allow the choice of stable, reliable although simple and

For what concerns the electronic design, the two main topics are the miniaturization of the circuitry and the reduction of power consumption in order to reach at least 5 days of continuous working with small, commercial batteries Our technological partners already designed and developed products or prototypes able to collect the proposed signals through wearable sensors in a reliable way, but they need to be improved since all the solutions are not optimized in terms of power consumption and scalability The design of a SIP solution requires the refinement of previous solutions and the research of new components that can

be integrated with it The choice of new elements is one of the strategies taken to achieve a low-power design, even if it implies to test again the performance of the system in terms of signal quality, signal-to-noise ratio, drifts and all the parameters influencing the compliance with the medical specifications The final design, the list of components and the features of the system as output of the whole project will be released after the completion of the ICG tests To date, the logical structure of the prototype in use is described in Fig 3

Signals are real-time recorded and processed in order to extract parameters relevant for the clinicians who will receive them through the remote server Dialling, connection and authentication procedures are directly controlled by the wearable unit, thus excluding any user intervention The processing output is directly stored in the remote database The use

of a new generation of 32-bit microcontroller unit (MCU) allows the management of the entire process, optimizing the power consumption at the same time

Different strategies of power management are investigated for each working condition

In fact there are two different situations:

- the normal one, when only brief parameters are transmitted, once in every minute;

- the “alarm” condition, during which also the raw signals are transmitted

This latter configuration allows a prompt analysis of the ECG by the clinicians, who can decide to reject the alarm or to activate countermeasures

This fully automated model of service is based on the possibility to identify different critical conditions from the biosignals applying the rules provided by the clinicians and embedded

in the CPU After the trigger of a possible critical situation has launched, the DUN Bluetooth profile tries to directly connect itself to the remote server and to transmit the raw signals, beginning from the past last minute, until the remote operator will decide to stop the “alarm condition” Its management implies the possibility to use also GPRS data transmission, because the 3G network coverage is still not ensured in all the neighbouring areas As a consequence, we are also exploring data compression algorithms for raw signals transmission in case of poor mobile network coverage

Trang 12

Fig 3 The electronic device, the SIP components are evidenced

SERVER and WEB-SERVICE

The hospital security model generally denies the possibility to directly connect a remote device to a local server, physically and logically placed inside the hospital, according to both the Italian and other countries regulation On the same purpose, it is worth noting that it is not mandatory placing the main data collector inside the hospital network, in fact through a secure web-service application the operators can access data while being in the hospital, or more precisely, in a control room where the physiological parameters and alarms can be constantly observed by them The use of different servers in order to cover the involved area

in a better way and to create a local node to distribute the collision management overhead in case of multiple alarm conditions is thus made possible The problem of priority management could not be faced at the mobile phone level because a typical mobile service provider can grant dedicated server resources only with business contracts The mobile phone is just a transparent modem connected to the server The policy in case of a coincidence of alarms is a matter of debate, but speaking from the informatics point of view,

as soon as the connection mobile-server is established through the standard port, the link should be automatically turned on a specific port, in order to set free the common resource for the next alarm

BT Module

3-AXIS Accelerometer

SIP

Trang 13

Fig 3 The electronic device, the SIP components are evidenced

SERVER and WEB-SERVICE

The hospital security model generally denies the possibility to directly connect a remote

device to a local server, physically and logically placed inside the hospital, according to both

the Italian and other countries regulation On the same purpose, it is worth noting that it is

not mandatory placing the main data collector inside the hospital network, in fact through a

secure web-service application the operators can access data while being in the hospital, or

more precisely, in a control room where the physiological parameters and alarms can be

constantly observed by them The use of different servers in order to cover the involved area

in a better way and to create a local node to distribute the collision management overhead in

case of multiple alarm conditions is thus made possible The problem of priority

management could not be faced at the mobile phone level because a typical mobile service

provider can grant dedicated server resources only with business contracts The mobile

phone is just a transparent modem connected to the server The policy in case of a

coincidence of alarms is a matter of debate, but speaking from the informatics point of view,

as soon as the connection mobile-server is established through the standard port, the link

should be automatically turned on a specific port, in order to set free the common resource

for the next alarm

Bat

BT Module

3-AXIS Accelerometer

SIP

4 Preliminary results and conclusions

During one year and half of work the consortium faced the problems related to the industrialisation and certification of the product Since some partners were already involved

in such area of the market, preliminary operations like market survey or patent analysis were almost ready at the early stage of the project This way it was possible to promptly release the main specifications for the system, although the industrialisation process is still a work in progress In fact the targeted customers population, predominantly composed by not-or-less technology skilled people, requires a detailed analysis on specific components responsible for the usability and allowing a user-friendly system’s management and handling

As anticipated, we are currently addressing the testing phase on the t-shirts in terms of biocompatibility and clinical performances

Regarding the hardware, we are testing the low-scaled device The SIP will be the final result of the project because the design and production of a SIP system is an expensive and complex process, requiring a lot of efforts in order to reduce the risk of a major fault For this reason we are carrying out specific tests collecting data on performances and trying to understand potential criticism before starting the first production

Current status Industrialisation

problems specifications [%] Compliance with

validating 2 nd release Sensors embedding and connectors 70%

T-shirts 100%

consolidation Refine power management 80%

Table 1 Current project checkout list

During the last months, the firmware has been tested on the same devices (MCU and Bluetooth module) in terms of power consumption and transmission throughput Based on current measures, we forecast it will be possible to ensure a 5-days working time with a commercial Lithium-Ion or Lithium-Polymer battery with less than 500mAh of capacity with the transmission protocol already working

Although there are still some critical aspects highlighted in Table 1, several problems related

to the creation of a commercial unobtrusive and fully automated wearable monitoring solution have been solved Moreover a great boost to the project has come from the introduction of a System in Package solution, the heart of the electronic device, which could probably have a deep impact also in next products and projects

Trang 14

5 References

Andreoni G (2008), Sistemi di sensori indossabili per il monitoraggio: Dalla Ricerca al Mercato, In:

Bonfiglio A., Cerutti S., De Rossi D., Magenes G (eds.), Sistemi Indossabili Intelligenti per la salute e la protezione dell’uomo, Patron, 2008

COM 689 (2008) Communication from the commission to the European Parliament, the

Council, the European Economic and Social Committes and the Committee of the Regions on telemedicine for the benefit of patients, healthcare systems and society,

COMMISSION OF THE EUROPEAN COMMUNITIES, Brussels

David K (2007), Wearable Electronics Systems Global Market Demand Analysis: Health Care

Solutions, In: VDC Research Report # VDC6520

Lymberis A and De Rossi D (2004) Wearable eHealth Systems for Personalised Health

Management State of the Art and Future Challenges, IOS Press, ISBN I 58603 449 9,

Netherlands

Lymberis A., Gatzoulis L (2006), Wearable Health Systems: from smart technologies to real

applications Conf Proc IEEE Eng Med Biol Soc.: 6789-92

Lymberis A and Paradiso R (2008) Smart Fabrics and Interactive Textile Enabling

Wearable Personal Applications: R&D State of the Art and Future Challenges,

Proceedings of 30th Annual International IEEE EMBS Conference, pp 5270-5273,

Vancouver, British Columbia, August 2008, Canada

Pantepopulous A and Bourbakis N (2008) A Survey on Wearable Biosensor Systems for

Health Monitoring, Proceedings of 30th Annual International IEEE EMBS Conference,

pp 4887-4890, Vancouver, British Columbia, August 2008, Canada

Trang 15

Neuro-Developmental Engineering: towards early diagnosis of neuro-developmental disorders 685

Neuro-Developmental Engineering: towards early diagnosis of developmental disorders

neuro-Domenico Campolo, Fabrizio Taffoni, Giuseppina Schiavone, Domenico Formica, Eugenio Guglielmelli and Flavio Keller

0

Neuro-Developmental Engineering: towards early

diagnosis of neuro-developmental disorders

1Domenico Campolo, Fabrizio Taffoni, Giuseppina Schiavone, Domenico Formica, Eugenio Guglielmelli and Flavio Keller

Università Campus Bio-Medico

00128 Roma - Italy

Nanyang Technological University

639798 Singapore

1 Introduction

Neuro-Developmental Engineering (NDE) is a new and emerging interdisciplinary research

area at the intersection of developmental neuroscience and bioengineering aiming at

provid-ing new methods and tools for: i) understanding neuro-biological mechanisms of human

brain development; ii)quantitative analysis and modeling of human behavior during

neuro-development; iii)assessment of neuro-developmental milestones achieved by humans from

birth onwards

Main application fields of NDE are:

- New clinical protocols and standards for early diagnosis, functional evaluation and

therapeutic treatments of neuro- developmental disorders;

- New generations of educational, interactive toys which can provide adequate stimuli

and guidance for supporting the physiological neuro-development process

This technology is expected to be also useful in the long term for developing new tools, e.g

toys, which can sustain, in ecological scenarios, the regular development of motor and

cogni-tive abilities of the child, based on a rigorous scientific approach

The long term goal is establishing standards against which development of infants at risk

for neuro-developmental disorders, particularly autism, can be measured, with the aim of

detecting early signs of disturbed development

1.1 Sensori-Motor Integration Deficits in Neurodevelopmental Disorders

Neurodevelopmental disorders such as ASD, ADHD, Tourette syndrome and others are

char-acterized by a genetic basis In this case behavioral analysis, or behavioral phenotyping, will

be instrumental for the analysis of the roles of genes in behavior (Gerlai 2002)

Autism is a behavioral disorder, with onset in childhood, which is characterized by deficits in

three basic domains: social interaction, language and communication, and pattern of interests

There is no doubt that autism has a strong genetic component, and that biological disease

mechanisms leading to autism are already active during foetal development and/or infancy,

35

Trang 16

as demonstrated, for example, by the abnormal pattern of brain growth during late foetal and

early postnatal life (see (Keller and Persico 2003), for a review) Autism is usually diagnosed

at the age of 3 years, in many cases after a period of seemingly normal neurological and

behavioral development The diagnosis of autism is purely clinical, there are no laboratory

tests to confirm or disprove the diagnosis It has been recognized that, although typical autism

is not associated with major neurological deficits, autism has characteristic manifestations in

the perceptual and motor domains.

Deficits in the perceptual domain include altered processing and recognition of socially relevant

information from peopleŠs faces (see (Grelotti et al 2003), for a review), deficits in

percep-tion of mopercep-tion cues (Milne et al 2002), (Spencer et al 2000), (Bertone et al 2003), (Takerae

et al 2004), difficulty in disengaging attention (Landry and Bryson 2004) and alterations of

auditory processing (Courchesne et al 1984), (Boddaert et al 2004) Studies based on

analy-sis of home-made movies suggest that an impairment of spontaneous attention toward social

stimuli is present already at 20 months (Swettenham et al 1998), and possibly also as early as

during the first 6 months of life (Maestro et al 2002) Furthermore, an autism-like syndrome is

frequently observed in congenitally blind children (Hobson and Bishop 2003) Taken together,

these observations suggest that at least some individuals with autism are characterized by an

early deficit of ‘low-level’ perceptual processing, which jeopardizes their ability to develop

higher-level capacities, such as language and interpersonal skills

Motor impairments in autism include deficits in postural reflexes (Minshew et al 2004),

(Schmitz et al 2003), (Molloy et al 2003), repetitive, stereotyped movements and awkward

patterns of object manipulation, lack of purposeful exploratory movements (see e.g (Pierce

and Courchesne 2001)), gaze abnormalities (Sweeney et al 2004), unusual gait pattern (Hallett

et al 1993), and alterations of movement planning and execution, which express themselves

as ‘hyper- dexterity’ (Rinehart et al 2001), (Mari et al 2003) Motor abnormalities may be

observed retrospectively in infants who later develop the autistic syndrome, on the basis of

home-made movies made during the first year of life (Teitelbaum et al 1998), (Teitelbaum et

al 2004) These clinical observations are consistent with a large body of evidence of subtle

structural and functional abnormalities of cortical and subcortical neural systems involved

in movement planning and execution, such as the prefrontal cortex, the basal ganglia and the

cerebellum (see (Keller and Persico 2003), for a review)

1.2 Ecological Approach

The diagnosis of ASD is currently made at 3 years of age; Attention-Deficit Hyperactivity

Dis-order (ADHD) is always considered as an alternative diagnosis of “high functioning” autism;

Tourette syndrome is diagnosed at age 7 or later ASD is therefore a natural candidate for

demonstrating the validity of novel approaches to early diagnosis As shown in Fig 1,

in-fancy, i.e the first 2-3 years of life before language development, represents an important

temporal window for an early diagnosis of ASD

The goal of our approach is twofold On one hand, guided by neuroscientists, we develop

technological platforms and methods to extract more information on perceptual and

intersub-jective capacities of human infants than is currently possible; this information could be later

used for early diagnosis of developmental disorders On the other hand, infancy provides

us with an important window of opportunity to capture the mechanisms behind

sensorimo-tor integration as these are just developing Moreover, neurodevelopmental disorders are an

important benchmark to highlight failures within such mechanism Such a knowledge can

be useful to neuroscientists to better understand the human brain functions involved in the

age (years)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

Autism ADHD Tourette Syndrome

infancy

current diagnostic tools

window of opportunity

brain development

Fig 1 Current diagnostic tools

sensorimotor integration but also to engineers, providing unique insights on how to buildcomplex and adaptable artificial systems (Metta et al 1999)

2 Technology for Assessing Movement and Gaze 2.1 Motion Tracking

Motion tracking can count on a host of different technological solutions, operating on entirelydifferent physical principles, with different performance characteristics and designed for dif-ferent purposes As shown in (Welch and Foxlin 2002), there is not a single technology thatcan fit all needs Each application defines the best technology to be implemented In order toperform a selection, the main characteristics of available technologies are briefly summarizedhereafter (see (Welch and Foxlin 2002) for details)

Mechanical sensing:

typically used for body motion capture; it uses angle and range measurements with the help

of gears and bend sensors; very accurate but bulky, often limiting mobility.

Optical sensing:

several principles are available, typical systems are camera- based ones; position of markers

in 3D space can be estimated very accurately within working volume (typically a few cubemeters, depending on the number of deployed cameras); line-of- sight issues (i.e the fact thatbody parts or other objects may occlude the visual scene of a camera, losing thus the sight

of one or more markers) is a limiting factor; very expensive; often requires highly structuredenvironments, at least when high accuracy is needed

Acoustic sensing:

typically based on time-of-flight of ultrasound pulses between emitters and receivers; sound

speed in air (about 340 m/s, resulting in sampling periods in the order of a few tens of

mil-liseconds) is slow but still acceptable for sensing human ( in particular infants) movements;line-of-sight issues are not as severe as for the optical technology; requires much less struc-tured environments than optical trackers; suitable to be used in ecological conditions (e.g.kindergartens)

Trang 17

Neuro-Developmental Engineering: towards early diagnosis of neuro-developmental disorders 687

as demonstrated, for example, by the abnormal pattern of brain growth during late foetal and

early postnatal life (see (Keller and Persico 2003), for a review) Autism is usually diagnosed

at the age of 3 years, in many cases after a period of seemingly normal neurological and

behavioral development The diagnosis of autism is purely clinical, there are no laboratory

tests to confirm or disprove the diagnosis It has been recognized that, although typical autism

is not associated with major neurological deficits, autism has characteristic manifestations in

the perceptual and motor domains.

Deficits in the perceptual domain include altered processing and recognition of socially relevant

information from peopleŠs faces (see (Grelotti et al 2003), for a review), deficits in

percep-tion of mopercep-tion cues (Milne et al 2002), (Spencer et al 2000), (Bertone et al 2003), (Takerae

et al 2004), difficulty in disengaging attention (Landry and Bryson 2004) and alterations of

auditory processing (Courchesne et al 1984), (Boddaert et al 2004) Studies based on

analy-sis of home-made movies suggest that an impairment of spontaneous attention toward social

stimuli is present already at 20 months (Swettenham et al 1998), and possibly also as early as

during the first 6 months of life (Maestro et al 2002) Furthermore, an autism-like syndrome is

frequently observed in congenitally blind children (Hobson and Bishop 2003) Taken together,

these observations suggest that at least some individuals with autism are characterized by an

early deficit of ‘low-level’ perceptual processing, which jeopardizes their ability to develop

higher-level capacities, such as language and interpersonal skills

Motor impairments in autism include deficits in postural reflexes (Minshew et al 2004),

(Schmitz et al 2003), (Molloy et al 2003), repetitive, stereotyped movements and awkward

patterns of object manipulation, lack of purposeful exploratory movements (see e.g (Pierce

and Courchesne 2001)), gaze abnormalities (Sweeney et al 2004), unusual gait pattern (Hallett

et al 1993), and alterations of movement planning and execution, which express themselves

as ‘hyper- dexterity’ (Rinehart et al 2001), (Mari et al 2003) Motor abnormalities may be

observed retrospectively in infants who later develop the autistic syndrome, on the basis of

home-made movies made during the first year of life (Teitelbaum et al 1998), (Teitelbaum et

al 2004) These clinical observations are consistent with a large body of evidence of subtle

structural and functional abnormalities of cortical and subcortical neural systems involved

in movement planning and execution, such as the prefrontal cortex, the basal ganglia and the

cerebellum (see (Keller and Persico 2003), for a review)

1.2 Ecological Approach

The diagnosis of ASD is currently made at 3 years of age; Attention-Deficit Hyperactivity

Dis-order (ADHD) is always considered as an alternative diagnosis of “high functioning” autism;

Tourette syndrome is diagnosed at age 7 or later ASD is therefore a natural candidate for

demonstrating the validity of novel approaches to early diagnosis As shown in Fig 1,

in-fancy, i.e the first 2-3 years of life before language development, represents an important

temporal window for an early diagnosis of ASD

The goal of our approach is twofold On one hand, guided by neuroscientists, we develop

technological platforms and methods to extract more information on perceptual and

intersub-jective capacities of human infants than is currently possible; this information could be later

used for early diagnosis of developmental disorders On the other hand, infancy provides

us with an important window of opportunity to capture the mechanisms behind

sensorimo-tor integration as these are just developing Moreover, neurodevelopmental disorders are an

important benchmark to highlight failures within such mechanism Such a knowledge can

be useful to neuroscientists to better understand the human brain functions involved in the

age (years)

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

Autism ADHD Tourette Syndrome

infancy

current diagnostic tools

window of opportunity

brain development

Fig 1 Current diagnostic tools

sensorimotor integration but also to engineers, providing unique insights on how to buildcomplex and adaptable artificial systems (Metta et al 1999)

2 Technology for Assessing Movement and Gaze 2.1 Motion Tracking

Motion tracking can count on a host of different technological solutions, operating on entirelydifferent physical principles, with different performance characteristics and designed for dif-ferent purposes As shown in (Welch and Foxlin 2002), there is not a single technology thatcan fit all needs Each application defines the best technology to be implemented In order toperform a selection, the main characteristics of available technologies are briefly summarizedhereafter (see (Welch and Foxlin 2002) for details)

Mechanical sensing:

typically used for body motion capture; it uses angle and range measurements with the help

of gears and bend sensors; very accurate but bulky, often limiting mobility.

Optical sensing:

several principles are available, typical systems are camera- based ones; position of markers

in 3D space can be estimated very accurately within working volume (typically a few cubemeters, depending on the number of deployed cameras); line-of- sight issues (i.e the fact thatbody parts or other objects may occlude the visual scene of a camera, losing thus the sight

of one or more markers) is a limiting factor; very expensive; often requires highly structuredenvironments, at least when high accuracy is needed

Acoustic sensing:

typically based on time-of-flight of ultrasound pulses between emitters and receivers; sound

speed in air (about 340 m/s, resulting in sampling periods in the order of a few tens of

mil-liseconds) is slow but still acceptable for sensing human ( in particular infants) movements;line-of-sight issues are not as severe as for the optical technology; requires much less struc-tured environments than optical trackers; suitable to be used in ecological conditions (e.g.kindergartens)

Trang 18

(Geo)Magnetic sensing:

a first method is based on electromagnetic coupling between a source and several trackers;

main drawbacks are that signal decays as 1/d3 (where d is the source-tracker distance) and

is affected by the geomagnetic field; these devices are quite expensive and require a certain

amount of structuring of the environment A second method is electronic compassing;

esti-mates heading and solely relies on the geomagnetic field, i.e it does not require any artificial

source and is therefore sourceless; measurements can be altered by ferromagnetic influence of

surrounding objects

Inertial sensing:

highly miniaturized accelerometers and gyroscopes are used to sense, respectively,

acceler-ation (comprising the gravitacceler-ational field) and angular velocity; used as inclinometers,

ac-celerometers can sense the gravity vector, i.e the ‘vertical’ direction, in this sense they are

also sourceless.

temperature fluctuations

position orientation

working volume cost

ferromagnetic influence acoustic noise optical noise

Inertial Geomagnetic Acoustic Optical Mechanical

+

- -

+- +-

- -

LEGEND:

tracker size / obstructive

Fig 2 Selection chart of different motion tracking technologies

In Fig 2, a selection chart for the different available technologies is provided For each

avail-able technology (columns) its suitability with respect to the performance characteristics of

interest (rows) is indicated Since our main purpose is developing technological tools that are

either wearable by infants or embeddable into toys, the highest priority is given to

technolo-gies which are unobtrusive This directly leads us to discard solutions involving mechanical

trackers

The second element considered for selection are the line-of-sight issues, since we are going to

deal with infants, it is extremely difficult to perform experiments with technologies that are

limited by the line of sight, a peculiarity of the optical technology which is only suitable to

experiments with collaborative subjects who are somehow willing to ‘act’ in front of a era Line-of-sight issues are much less severe for the acoustic technology which is thus stillappealing for movements analysis in infants

cam-The third element of the selection criterion is performance with respect to tracking accuracy.Here a distinction is made between tracking positions and tracking orientations Measure-ment principles such as the time-of-flight (typically deployed in acoustic measurements) orcamera-based tracking are inherently suitable to measure the distance of points (markers) andthe origin of the measurement system (e.g the source of acoustic waves or a camera etc ).Orientations can be inferred indirectly by estimating distances between two or more mark-ers and the source of measurement The larger the distance between two markers, the betterthe estimation of orientation As dimensions shrink, as in the case of infants, accuracy ofindirect orientation measurements also decreases (e.g accurate tracking of the orientation

of an infant’s wrist can be problematic even without considering line-of-sight issues) Othertechnologies allow a direct measurement of orientations (for example inertial sensors used

as inclinometers can sense deviations from the vertical axis while magnetic sensors used ascompasses can sense deviations from the horizontal geomagnetic north direction) without re-quiring the positioning of multiple markers

As long as orientation is concerned, inertial and magnetic technologies appear to be very pealing since: are highly unobtrusive due the availability of miniaturized off-the-shelf devices;

ap-do not suffer from line-of-sight issues; can provide high accuracy in orientation tracking are

sourceless: do not require any structuring of the environment; have virtually unlimited

work-ing volume; are low-cost

The bottom half of Fig 2 shows, for each technology, the main limiting factors to a correctoperation Besides temperature, which affects any electrical device and that can be compen-sated in most of the cases, the real limiting factor for the magnetic technology is the presence

of ferromagnetic materials Common ferromagnetic objects such as iron parts of doors, chairs,tables etc can produce local distortions of the geomagnetic field, causing thus errors in theestimations of orientations As discussed in (Kemp et al 1998), some care should be taken,when conducting experiments, to avoid large ferromagnetic objects in the surroundings Wefound that this can be easily done in environments such as day-cares where, for safety reasons,all metals are usually avoided and typical materials used with children are wood, rubber andplastic

1 Magnetic Induction Method (Search Coil)

2 Electro-Oculography (EOG)

Trang 19

Neuro-Developmental Engineering: towards early diagnosis of neuro-developmental disorders 689

(Geo)Magnetic sensing:

a first method is based on electromagnetic coupling between a source and several trackers;

main drawbacks are that signal decays as 1/d3 (where d is the source-tracker distance) and

is affected by the geomagnetic field; these devices are quite expensive and require a certain

amount of structuring of the environment A second method is electronic compassing;

esti-mates heading and solely relies on the geomagnetic field, i.e it does not require any artificial

source and is therefore sourceless; measurements can be altered by ferromagnetic influence of

surrounding objects

Inertial sensing:

highly miniaturized accelerometers and gyroscopes are used to sense, respectively,

acceler-ation (comprising the gravitacceler-ational field) and angular velocity; used as inclinometers,

ac-celerometers can sense the gravity vector, i.e the ‘vertical’ direction, in this sense they are

also sourceless.

temperature fluctuations

position orientation

working volume cost

ferromagnetic influence acoustic noise

+

- -

+-

-

-

LEGEND:

tracker size / obstructive

Fig 2 Selection chart of different motion tracking technologies

In Fig 2, a selection chart for the different available technologies is provided For each

avail-able technology (columns) its suitability with respect to the performance characteristics of

interest (rows) is indicated Since our main purpose is developing technological tools that are

either wearable by infants or embeddable into toys, the highest priority is given to

technolo-gies which are unobtrusive This directly leads us to discard solutions involving mechanical

trackers

The second element considered for selection are the line-of-sight issues, since we are going to

deal with infants, it is extremely difficult to perform experiments with technologies that are

limited by the line of sight, a peculiarity of the optical technology which is only suitable to

experiments with collaborative subjects who are somehow willing to ‘act’ in front of a era Line-of-sight issues are much less severe for the acoustic technology which is thus stillappealing for movements analysis in infants

cam-The third element of the selection criterion is performance with respect to tracking accuracy.Here a distinction is made between tracking positions and tracking orientations Measure-ment principles such as the time-of-flight (typically deployed in acoustic measurements) orcamera-based tracking are inherently suitable to measure the distance of points (markers) andthe origin of the measurement system (e.g the source of acoustic waves or a camera etc ).Orientations can be inferred indirectly by estimating distances between two or more mark-ers and the source of measurement The larger the distance between two markers, the betterthe estimation of orientation As dimensions shrink, as in the case of infants, accuracy ofindirect orientation measurements also decreases (e.g accurate tracking of the orientation

of an infant’s wrist can be problematic even without considering line-of-sight issues) Othertechnologies allow a direct measurement of orientations (for example inertial sensors used

as inclinometers can sense deviations from the vertical axis while magnetic sensors used ascompasses can sense deviations from the horizontal geomagnetic north direction) without re-quiring the positioning of multiple markers

As long as orientation is concerned, inertial and magnetic technologies appear to be very pealing since: are highly unobtrusive due the availability of miniaturized off-the-shelf devices;

ap-do not suffer from line-of-sight issues; can provide high accuracy in orientation tracking are

sourceless: do not require any structuring of the environment; have virtually unlimited

work-ing volume; are low-cost

The bottom half of Fig 2 shows, for each technology, the main limiting factors to a correctoperation Besides temperature, which affects any electrical device and that can be compen-sated in most of the cases, the real limiting factor for the magnetic technology is the presence

of ferromagnetic materials Common ferromagnetic objects such as iron parts of doors, chairs,tables etc can produce local distortions of the geomagnetic field, causing thus errors in theestimations of orientations As discussed in (Kemp et al 1998), some care should be taken,when conducting experiments, to avoid large ferromagnetic objects in the surroundings Wefound that this can be easily done in environments such as day-cares where, for safety reasons,all metals are usually avoided and typical materials used with children are wood, rubber andplastic

1 Magnetic Induction Method (Search Coil)

2 Electro-Oculography (EOG)

Trang 20

3 Photoelectric Methods: Infra-Red (IR) Oculography

4 Video-Oculography (VOG)

Each methodology is characterized by parameters such as range of measurement, sensitivity,

linearity, accuracy, discomfort for the subject, interference with the field of view of the subject,

tolerance to head movement

Magnetic Induction Method (Search Coil):

The search coil technique has become the accepted standard for the measurement of 3D eye

movement This technique is based on the fact that a magnetic field induces a voltage in a coil

(search coil) which is attached to the eye The induced voltage has amplitude proportional to

the sine of the angle between the coil axis and the magnetic field direction The magnetic field

is provide by coils mounted at the sides of a cubic frame The dimensions of the sides of the

frame can vary from few tens of centimeters to few meters, allowing to measure also other

movements (i.e eye-hand coordination) Robinson (Robinson 1963) was the first to apply this

technique, using a coil secured to the eye by suction Nowadays the search coil is embedded

in a scleral contact lens The lens is subject to slippage if the lens covers only the cornea Eye

movement is measured in absolute spatial coordinates Head orientation can also be measured

with a search coil mounted on the forehead, and orientation and movement of the eye within

the head can be calculated from the orientation of the head and of the eye with respect to

the magnetic field (Haslwanter 1995) Currently a number of different search coil systems are

commercially available (e.g by Skalar Instruments, C-N-C Engineering, Remmel Labs, etc.)

Although the scleral search coil is the most precise eye movement measurement method (very

high temporal and spatial resolution can be obtained with accuracy to about 5-10 arc-seconds

over a limited range of about 5 deg), it is also the most intrusive method Insertion of the lens

requires care and practice and wearing the lens causes discomfort and risk of corneal abrasion

or lead breakage The requirements to stay in the center of the magnetic field precludes the

use of search coils during many natural activities Thus, this technique is mostly used for

research purposes, it is not suitable for clinical routine

Electro-Oculography:

First applications of electro-oculography are dated back to the ‘30s and are currently widely

used both for clinical and research purposes It relies on measurement of electrical potential

differences between the cornea and the retina, discovered by DuBois-Reymond in 1849 Skin

electrodes are positioned around the eye The measured potential difference is proportional to

the sine of the rotation angle of the eye For small rotation the proportionality is almost linear;

it decreases for higher angles of rotation (Byford 1963) The recorded potentials are in the

range 15-200 µV, with nominal sensitivities of order of 20 µV/deg of eye movement The eye

movement is measured in craniotopic coordinates and head movement during recording does

not affect the measurement The discomfort for the subject is limited and the measurement

range is wide both for horizontal movements (±70 deg) and for vertical movements (±30 deg),

even if the sensitivity decreases for lateral position of the eye The most important advantage

of this methodology is the possibility of recording eye movement with closed eyes, which is

relevant requirement during some experimental protocol (e.g during sleeping phases) The

main drawback of this technique are related to the nature of the potential recorded and to

the artifacts due to the electrodes properties As concern the potential, the resting

corneo-retinic potential (usually of the order of 0.4-1 mV) can be affected by lighting conditions of

the environment and by the psycho-physical condition of the subject The artifacts at the

level of the skin electrodes relies on the contact resistance electrode-skin, on the oxidation andpolarization of the electrodes

Infra-Red Oculography:

Infra-red (IR) oculography is based on the recording of the light reflected by the eye when

it is lighted with IR light beam Since IR light is not visible, it does not interfere with thesubject vision, moreover the IR detectors are not influenced by environmental lighting con-ditions There are three categories of Infra-red (IR) oculography which use respectively: thecorneal reflection, the Purkinje images and the track of the limb Due to the construction ofthe eye, when a beam of IR light points to it, four reflections are formed on the eye, calledPurkinje images (Cornsweet and Crane 1973): the first on the front surface of the cornea and

it is called corneal reflection, the second image on rear surface of the cornea, the third on thefront surface of the lens and the fourth on the rear surface of the lens By detecting the cornealreflection and the pupil center and by using an appropriate calibration procedure, it is possi-ble to measure the Point of Regard (gaze) on a planar surface on which calibration points arepositioned Two points of reference on the eye are needed to separate eye movements fromhead movements The positional difference between the pupil center and corneal reflectionchanges with pure eye rotation, but remains relatively constant with minor head movements.The corneal reflection moves in the opposite direction of the eye respect to the pupil center

In other cases both the first and the fourth Purkinje images (Dual-Purkinje images eye ers are detected Both reflections move together through exactly the same distance upon eyetranslation but they move through different distances upon eye rotation The third methodbased on photoelectric principle relies on the track of the limb (scleral-iris edge) of the eye bymeasuring the amount of scattered light Most photoelectric systems must be mounted close

track-to the eyes (i.e EL-MAR tracking device), so they may restrict the field of view, moreover fastmovements of the head can cause slippage of the device on the head leading to mis-alignment

of the eye respect to the IR emitter and detector There exist also external device and a supportfor keeping the head fixed is needed (i.e Tobii eye tracker) The range of measurement of thephotoelectric eye tracker is not higher then±30 deg in the horizontal plane and±20 deg inthe vertical plane

Video-Oculography:

Video systems for measuring ocular movements are based on the analysis of images recorded

by cameras This technique, introduced in the ‘80s, quickly improved in terms of mances and reliability thanks to the technological development of digital cameras and com-puter powerful The Video Oculography (VOG) provides directly a digital output Severalalgorithms are available for the pupil detection in an image frame and pupil centroid coor-dinates extraction, nevertheless environmental lighting conditions can affect the automaticdetection (Eizenman et al 1984) (Landau 1987) Thus, IR light is used together with videorecording, so that the pupil appears brighter This technique is called Pupil Center/CornealReflection (PC/CR) because the IR light produces also the Purkinje images, mentioned before

perfor-As in IR Oculography, also VOG can be realized both as wearable device (DiScenna 1995) andprovides measurements in craniotopic coordinates or external device and provides measure-ments in spatial coordinates Head-mounted system (i.e EyeLink) can be worn without toomuch discomfort High resolution and high frame rate CCD and CMOS cameras are used.Reduced dimensions and weight of the actual cameras allow to position them in such a waythat they interfere as less as possible with the field of view of the subject (Babcock and Pelz

Ngày đăng: 21/06/2014, 19:20