Pe r A n d e rs R ic ka rd H e lls trö m W EAR AB LE P EDO BARO G R AP H Y S YS TE M F O R M O N IT O R IN G O F W A LK R EL A TE D P AR AM ET ER S 2 019 ISBN 978-91-7485-444-2 ISSN 1651-4238
Address: P.O. Box 883, SE-721 23 Västerås. Sweden Address: P.O. Box 325, SE-631 05 Eskilstuna. Sweden E-mail: info@mdh.se Web: www.mdh.se
Wearable Pedobarography System for
Monitoring of Walk Related Parameters
Per Anders Rickard HellströmMälardalen University Doctoral Dissertation 301
Per Hellström received a Master of Science in electrical engineering in 2009, from the Royal Institute of Technology (KTH) in Stockholm, with a specialization in medical measurements and signal processing. He has working experience as a measurement technician in the automotive industry. His research focus is on biomedical sensor systems and he received a licentiate degree in electronics in 2016 from Mälardalen University (MDH) in Västerås. Naiara Seara created the illustration on the front cover.
Mälardalen University Press Dissertations No. 301
WEARABLE PEDOBAROGRAPHY SYSTEM FOR
MONITORING OF WALK RELATED PARAMETERS
Per Anders Rickard Hellström
2019
School of Innovation, Design and Engineering
Mälardalen University Press Dissertations No. 301
WEARABLE PEDOBAROGRAPHY SYSTEM FOR
MONITORING OF WALK RELATED PARAMETERS
Per Anders Rickard Hellström
2019
School of Innovation, Design and Engineering
Copyright © Per Anders Rickard Hellström, 2019 ISBN 978-91-7485-444-2
ISSN 1651-4238
Printed by E-Print AB, Stockholm, Sweden
Copyright © Per Anders Rickard Hellström, 2019 ISBN 978-91-7485-444-2
ISSN 1651-4238
Printed by E-Print AB, Stockholm, Sweden
Mälardalen University Press Dissertations No. 301
WEARABLE PEDOBAROGRAPHY SYSTEM FOR MONITORING OF WALK RELATED PARAMETERS
Per Anders Rickard Hellström
Akademisk avhandling
som för avläggande av teknologie doktorsexamen i elektronik vid Akademin för innovation, design och teknik kommer att offentligen försvaras fredagen
den 6 december 2019, 13.30 i Delta, Mälardalens högskola, Västerås. Fakultetsopponent: Docent Helena Grip, Umeå universitet
Akademin för innovation, design och teknik
Mälardalen University Press Dissertations No. 301
WEARABLE PEDOBAROGRAPHY SYSTEM FOR MONITORING OF WALK RELATED PARAMETERS
Per Anders Rickard Hellström
Akademisk avhandling
som för avläggande av teknologie doktorsexamen i elektronik vid Akademin för innovation, design och teknik kommer att offentligen försvaras fredagen
den 6 december 2019, 13.30 i Delta, Mälardalens högskola, Västerås. Fakultetsopponent: Docent Helena Grip, Umeå universitet
Akademin för innovation, design och teknik
Abstract
Health care costs have increased over the last decades due to an ageing population. Therefore, research in personal health monitoring (PHM) has increased in response to this. PHM has advantages such as mobility (monitoring of health at work or at home), early detection of health problems enabling preventive health measures and a reduction of health care cost. Human motion analysis, using for example pedobarography (PBG), is an important subcategory of PHM. PBG is used to study the force fields acting between the plantar surface of the foot and a supporting surface. Gait and posture analysis, prosthetics evaluation and monitoring of recovery from injury or disease are examples of PBG applications. Portable PBG can be performed using force sensing resistors built into the insole inside the shoe.
In accordance with this, the research aim for this thesis is to design, build and evaluate a wireless wearable measurement system based on PBG for monitoring of walk related parameters. Monitoring of carried weight and walking speed were chosen as the applications for validation of the system. Motivations for choosing these applications are that there is a lack of a wearable system for monitoring of weight while walking and a possible combination with accelerometers to improve the estimation of walking speed. Both walking speed and weight are important factors for estimating energy expenditure. A portable system, that estimates weight while walking, enables monitoring of heavy working conditions.
The main research contributions include design of a PBG measurement system with a sensor implementation resulting in good sensor durability, several novel methods for weight estimation during walk and a novel method for analysing walking intensity and relating it to walking speed. The research results show that the new PBG system, in combination with the novel analysing methods, are suitable for use in wearable systems for monitoring of health related walk parameters.
ISBN 978-91-7485-444-2 ISSN 1651-4238
Abstract
Health care costs have increased over the last decades due to an ageing population. Therefore, research in personal health monitoring (PHM) has increased in response to this. PHM has advantages such as mobility (monitoring of health at work or at home), early detection of health problems enabling preventive health measures and a reduction of health care cost. Human motion analysis, using for example pedobarography (PBG), is an important subcategory of PHM. PBG is used to study the force fields acting between the plantar surface of the foot and a supporting surface. Gait and posture analysis, prosthetics evaluation and monitoring of recovery from injury or disease are examples of PBG applications. Portable PBG can be performed using force sensing resistors built into the insole inside the shoe.
In accordance with this, the research aim for this thesis is to design, build and evaluate a wireless wearable measurement system based on PBG for monitoring of walk related parameters. Monitoring of carried weight and walking speed were chosen as the applications for validation of the system. Motivations for choosing these applications are that there is a lack of a wearable system for monitoring of weight while walking and a possible combination with accelerometers to improve the estimation of walking speed. Both walking speed and weight are important factors for estimating energy expenditure. A portable system, that estimates weight while walking, enables monitoring of heavy working conditions.
The main research contributions include design of a PBG measurement system with a sensor implementation resulting in good sensor durability, several novel methods for weight estimation during walk and a novel method for analysing walking intensity and relating it to walking speed. The research results show that the new PBG system, in combination with the novel analysing methods, are suitable for use in wearable systems for monitoring of health related walk parameters.
ISBN 978-91-7485-444-2 ISSN 1651-4238
Abstract
Health care costs have increased over the last decades due to an ageing population. Therefore, research in personal health monitoring (PHM) has increased in response to this. PHM has advantages such as mobility (monitoring of health at work or at home), early detection of health prob-lems enabling preventive health measures and a reduction of health care cost. Human motion analysis, using for example pedobarography (PBG), is an important subcategory of PHM. PBG is used to study the force fields acting between the plantar surface of the foot and a supporting surface. Gait and posture analysis, prosthetics evaluation and monitoring of recovery from injury or disease are examples of PBG applications. Portable PBG can be performed using force sensing resistors built into the insole inside the shoe.
In accordance with this, the research aim for this thesis is to design, build and evaluate a wireless wearable measurement system based on PBG for monitoring of walk related parameters. Monitoring of carried weight and walking speed were chosen as the applications for validation of the system. Motivations for choosing these applications are that there is a lack of a wearable system for monitoring of weight while walking and a possible combination with accelerometers to improve the estimation of walking speed. Both walking speed and weight are important factors for estimating energy expenditure. A portable system, that estimates weight while walking, enables monitoring of heavy working conditions.
The main research contributions include design of a PBG measurement system with a sensor implementation resulting in good sensor durability, several novel methods for weight estimation during walk and a novel method for analysing walking intensity and relating it to walking speed. The research results show that the new PBG system, in combination with the novel analysing methods, are suitable for use in wearable systems for monitoring of health related walk parameters.
Abstract
Health care costs have increased over the last decades due to an ageing population. Therefore, research in personal health monitoring (PHM) has increased in response to this. PHM has advantages such as mobility (monitoring of health at work or at home), early detection of health prob-lems enabling preventive health measures and a reduction of health care cost. Human motion analysis, using for example pedobarography (PBG), is an important subcategory of PHM. PBG is used to study the force fields acting between the plantar surface of the foot and a supporting surface. Gait and posture analysis, prosthetics evaluation and monitoring of recovery from injury or disease are examples of PBG applications. Portable PBG can be performed using force sensing resistors built into the insole inside the shoe.
In accordance with this, the research aim for this thesis is to design, build and evaluate a wireless wearable measurement system based on PBG for monitoring of walk related parameters. Monitoring of carried weight and walking speed were chosen as the applications for validation of the system. Motivations for choosing these applications are that there is a lack of a wearable system for monitoring of weight while walking and a possible combination with accelerometers to improve the estimation of walking speed. Both walking speed and weight are important factors for estimating energy expenditure. A portable system, that estimates weight while walking, enables monitoring of heavy working conditions.
The main research contributions include design of a PBG measurement system with a sensor implementation resulting in good sensor durability, several novel methods for weight estimation during walk and a novel method for analysing walking intensity and relating it to walking speed. The research results show that the new PBG system, in combination with the novel analysing methods, are suitable for use in wearable systems for monitoring of health related walk parameters.
Sammandrag
Kostnaderna f¨or sjukv˚ard har ¨okat de senaste decennierna p˚a grund av att vi lever allt l¨angre. D¨arf¨or s˚a har forskningen inom personlig h¨alsomonitorering (PHM) ¨okat som ett svar p˚a detta. PHM har f¨ordelar s˚asom mobilitet (monitorering av h¨alsa p˚a jobbet eller i hemmet), tidig uppt¨ackt av h¨alsoproblem g¨or det m¨ojligt att s¨atta in f¨orebyggande ˚atg¨arder f¨or h¨alsa och reducera kostnaden f¨or sjukv˚ard. R¨orelseanalys p˚a m¨anniskor, med hj¨alp av till exempel pedobarografi (PBG), ¨ar en viktig underkategori av PHM. PBG anv¨ands f¨or att studera kraftf¨alt som verkar mellan fotens undersida och en uppb¨arande yta. Analys av g˚angstil och kroppsh˚allning, utv¨ardering av proteser och ¨overvakning av ˚aterh¨amtning fr˚an skada eller sjukdom ¨ar exempel p˚a till¨ampningar f¨or
PBG.
I ¨overrensst¨ammelse med detta ¨ar syftet f¨or forskningen i den h¨ar avhand-lingen att utforma, bygga och utv¨ardera ett tr˚adl¨ost och b¨arbart m¨atsys-tem som bygger p˚a PBG f¨or ¨overvakning av g˚angrelaterade parametrar. ¨Overvakning av buren vikt och g˚anghastighet valdes som till¨ampningarna f¨or att utv¨ardera systemet. Motiveringar f¨or att v¨alja dessa till¨ampningar ¨ar att det finns en brist p˚a b¨arbara system f¨or ¨overvakning av vikt under g˚ang och att en m¨ojlig kombination med accelerometrar kan f¨orb¨attra uppskattningen av g˚anghastighet. B˚ade g˚anghastighet och vikt ¨ar viktiga faktorer vid uppskattning av energif¨orbrukning. Ett portabelt system, som uppskattar vikt under g˚ang, m¨ojligg¨or ¨overvakning av tunga arb-etsf¨orh˚allanden.
De fr¨amsta forskningsbidragen inkluderar utformningen av ett m¨atsystem baserat p˚a PBG med sensorimplementering som ger l˚ang livsl¨angd f¨or sensorerna, flera nya analysmetoder f¨or uppskattning av vikt under g˚ang och en ny analysmetod f¨or g˚angintensitet som relateras till g˚anghastighet. Forskningsresultaten visar p˚a att det nya PBG-systemet, i kombination med de nya analysmetoderna, ¨ar passande f¨or anv¨andning i b¨arbara system f¨or ¨overvakning av h¨alsorelaterade g˚angparametrar.
Sammandrag
Kostnaderna f¨or sjukv˚ard har ¨okat de senaste decennierna p˚a grund av att vi lever allt l¨angre. D¨arf¨or s˚a har forskningen inom personlig h¨alsomonitorering (PHM) ¨okat som ett svar p˚a detta. PHM har f¨ordelar s˚asom mobilitet (monitorering av h¨alsa p˚a jobbet eller i hemmet), tidig uppt¨ackt av h¨alsoproblem g¨or det m¨ojligt att s¨atta in f¨orebyggande ˚atg¨arder f¨or h¨alsa och reducera kostnaden f¨or sjukv˚ard. R¨orelseanalys p˚a m¨anniskor, med hj¨alp av till exempel pedobarografi (PBG), ¨ar en viktig underkategori av PHM. PBG anv¨ands f¨or att studera kraftf¨alt som verkar mellan fotens undersida och en uppb¨arande yta. Analys av g˚angstil och kroppsh˚allning, utv¨ardering av proteser och ¨overvakning av ˚aterh¨amtning fr˚an skada eller sjukdom ¨ar exempel p˚a till¨ampningar f¨or
PBG.
I ¨overrensst¨ammelse med detta ¨ar syftet f¨or forskningen i den h¨ar avhand-lingen att utforma, bygga och utv¨ardera ett tr˚adl¨ost och b¨arbart m¨atsys-tem som bygger p˚a PBG f¨or ¨overvakning av g˚angrelaterade parametrar. ¨Overvakning av buren vikt och g˚anghastighet valdes som till¨ampningarna f¨or att utv¨ardera systemet. Motiveringar f¨or att v¨alja dessa till¨ampningar ¨ar att det finns en brist p˚a b¨arbara system f¨or ¨overvakning av vikt under g˚ang och att en m¨ojlig kombination med accelerometrar kan f¨orb¨attra uppskattningen av g˚anghastighet. B˚ade g˚anghastighet och vikt ¨ar viktiga faktorer vid uppskattning av energif¨orbrukning. Ett portabelt system, som uppskattar vikt under g˚ang, m¨ojligg¨or ¨overvakning av tunga arb-etsf¨orh˚allanden.
De fr¨amsta forskningsbidragen inkluderar utformningen av ett m¨atsystem baserat p˚a PBG med sensorimplementering som ger l˚ang livsl¨angd f¨or sensorerna, flera nya analysmetoder f¨or uppskattning av vikt under g˚ang och en ny analysmetod f¨or g˚angintensitet som relateras till g˚anghastighet. Forskningsresultaten visar p˚a att det nya PBG-systemet, i kombination med de nya analysmetoderna, ¨ar passande f¨or anv¨andning i b¨arbara system f¨or ¨overvakning av h¨alsorelaterade g˚angparametrar.
To the love of my life, Noelia!
To the love of my life, Noelia!
Acknowledgement
First and foremost I would like to express my deepest gratitude to my main supervisor Mia Folke and my assistant supervisors Martin Ekstr¨om and Ning Xiong. They have given me large portions of their valuable time and taught me a lot, not only about the academic world and how to conduct research. I also owe much to my co-authors Anna ˚Akerberg, Lennie Carl´en Eriksson and Jonatan Scharff Willners. My writing skills have improved thanks to the thorough and thoughtful feedback I have received, especially from Anna.
The research presented in this doctoral thesis was conducted at the School of Innovation, Design and Technology (IDT), at M¨alardalen University in V¨aster˚as, Sweden, with financial support from the Swedish Knowledge Foundation (KKS). I am very grateful for their funding of the research profile Embedded Systems for Health (ESS-H).
I thank Maria Lind´en, research profile leader for ESS-H, for giving me this opportunity by selecting me from all of the applicants for this doctoral student position. I also would like to thank the rest of the ESS-H research group, Per Olov Risman, Maria Ehn, Mobyen Uddin Ahmed, Shaibal Barua, Shahina Begum, Mats Bj¨orkman, Aida Causevic, Jiaying Du, Mikael Ekstr¨om, Hossein Fotouhi, Peter Funk, Hamid GholamHosseini, Melika Hozhabri, Jonas Ljungblad, Jos´e Fern´an Mart´ınez Ortega, Ham-idur Rahman, Elisabeth Uhlemann, Maryam Vahabi, Ivan Tomasic, Johan ˚
Akerberg and Elaine ˚Astrand, for creating a very open and welcoming working environment.
I am part of the Intelligent Future Technology (IFT) division and I would like to thank Niklas Persson, Carl Ahlberg, Lars Asplund, Baran C¸¨ur¨ukl¨u, Fredrik Ekstrand, H˚akan Forsberg, Henrik Johansson, Miguel Leon Ortiz, Arash Ghareh Baghi, Farid Monsefi, Peter Ravenhold and Mirko Senkovski for wonderful discussions. Arash and Farid have really
Acknowledgement
First and foremost I would like to express my deepest gratitude to my main supervisor Mia Folke and my assistant supervisors Martin Ekstr¨om and Ning Xiong. They have given me large portions of their valuable time and taught me a lot, not only about the academic world and how to conduct research. I also owe much to my co-authors Anna ˚Akerberg, Lennie Carl´en Eriksson and Jonatan Scharff Willners. My writing skills have improved thanks to the thorough and thoughtful feedback I have received, especially from Anna.
The research presented in this doctoral thesis was conducted at the School of Innovation, Design and Technology (IDT), at M¨alardalen University in V¨aster˚as, Sweden, with financial support from the Swedish Knowledge Foundation (KKS). I am very grateful for their funding of the research profile Embedded Systems for Health (ESS-H).
I thank Maria Lind´en, research profile leader for ESS-H, for giving me this opportunity by selecting me from all of the applicants for this doctoral student position. I also would like to thank the rest of the ESS-H research group, Per Olov Risman, Maria Ehn, Mobyen Uddin Ahmed, Shaibal Barua, Shahina Begum, Mats Bj¨orkman, Aida Causevic, Jiaying Du, Mikael Ekstr¨om, Hossein Fotouhi, Peter Funk, Hamid GholamHosseini, Melika Hozhabri, Jonas Ljungblad, Jos´e Fern´an Mart´ınez Ortega, Ham-idur Rahman, Elisabeth Uhlemann, Maryam Vahabi, Ivan Tomasic, Johan ˚
Akerberg and Elaine ˚Astrand, for creating a very open and welcoming working environment.
I am part of the Intelligent Future Technology (IFT) division and I would like to thank Niklas Persson, Carl Ahlberg, Lars Asplund, Baran C¸¨ur¨ukl¨u, Fredrik Ekstrand, H˚akan Forsberg, Henrik Johansson, Miguel Leon Ortiz, Arash Ghareh Baghi, Farid Monsefi, Peter Ravenhold and Mirko Senkovski for wonderful discussions. Arash and Farid have really
Acknowledgement
opened my eyes for the Iranian culture. I would also love to thank all my wonderful current and past roommates, Masoud Daneshtalab, Jonatan Tidare, Johan Holmberg, Sara Abbaspour, Jimmie Hagblad, Vincent Hommersen, Gregory Koshmak, Elena Lisova, Kin Yun Lum, Branko Miloradovic, Mujtaba Aljumah, Mattias Olausson and Nikola Petrovic. It is a pleasure to share office with you.
LaTeX is an invaluable tool and I would like to thank Saad Mubeen, Ayhan Mehmed, Lars Hellstr¨om and Marina Guti´errez Lopez for inspiration about academic writing in LaTeX. I would also like to thank Moris Behnam, Svetlana Girs, Hans Hansson, Damir Isovic, Rikard Lindell, Anatoliy Malyarenko, Thomas Nolte, Sara Afshar, Mohammad Ashjaei, Lars Bark, Alessio Bucaioni, Simin Cai, Mirgita Frasheri, Pablo Guti´errez Pe´on, Leo Hatvani, Daniel Hedin, Linus K¨allberg, Ashalatha Kunnappilly, Bj¨orn Lisper, Filip Markovic, Dag Nystr¨om, Francisco Manuel Pozo P´erez, Apala Ray, Andreas Ryve, Antti Salonen, Irfan Sljivo, Jing Yue, Jiale Zhou and Guillermo Rodriguez Navas, and anyone else I may have forgotten to list here, for almost always having time for a conversation. Guillermo taught me about plantar fasciitis and how to alleviate the pain and it was very helpful.
I am, of course, deeply grateful to all the esteemed volunteers who participated in my walk experiments. Many of you also shared insightful comments and showed great interest in my research.
Our terrific administrators deserve a special thank you for their tireless work. Especially Annika Havbrandt, Natalie Caballero L¨of, Therese Jagestig Bjurquist, Carola Ryttersson, Susanne Fronn˚a, Malin ˚Ashuvud, Moa ¨Onell and Jenny H¨agglund have all been very helpful to me. Per Nystr¨om and Tord Heljeberg, from the library of M¨alardalen University, have provided highly valued research support.
Last, but not least, I owe a lot of gratitude to my dear family and friends for all their positivity and love. The greatly skilled Naiara Seara created an excellent illustration for the front cover. My eminent wife Noelia Gutierrez Rodriguez together with Anders, Ingrid, Julia, Joel and Aileen Hellstr¨om have always believed in me and I have received immense support from everyone. This journey would never have been possible without you!
Per Anders Rickard Hellstr¨om V¨aster˚as, October 2019
vi
Acknowledgement
opened my eyes for the Iranian culture. I would also love to thank all my wonderful current and past roommates, Masoud Daneshtalab, Jonatan Tidare, Johan Holmberg, Sara Abbaspour, Jimmie Hagblad, Vincent Hommersen, Gregory Koshmak, Elena Lisova, Kin Yun Lum, Branko Miloradovic, Mujtaba Aljumah, Mattias Olausson and Nikola Petrovic. It is a pleasure to share office with you.
LaTeX is an invaluable tool and I would like to thank Saad Mubeen, Ayhan Mehmed, Lars Hellstr¨om and Marina Guti´errez Lopez for inspiration about academic writing in LaTeX. I would also like to thank Moris Behnam, Svetlana Girs, Hans Hansson, Damir Isovic, Rikard Lindell, Anatoliy Malyarenko, Thomas Nolte, Sara Afshar, Mohammad Ashjaei, Lars Bark, Alessio Bucaioni, Simin Cai, Mirgita Frasheri, Pablo Guti´errez Pe´on, Leo Hatvani, Daniel Hedin, Linus K¨allberg, Ashalatha Kunnappilly, Bj¨orn Lisper, Filip Markovic, Dag Nystr¨om, Francisco Manuel Pozo P´erez, Apala Ray, Andreas Ryve, Antti Salonen, Irfan Sljivo, Jing Yue, Jiale Zhou and Guillermo Rodriguez Navas, and anyone else I may have forgotten to list here, for almost always having time for a conversation. Guillermo taught me about plantar fasciitis and how to alleviate the pain and it was very helpful.
I am, of course, deeply grateful to all the esteemed volunteers who participated in my walk experiments. Many of you also shared insightful comments and showed great interest in my research.
Our terrific administrators deserve a special thank you for their tireless work. Especially Annika Havbrandt, Natalie Caballero L¨of, Therese Jagestig Bjurquist, Carola Ryttersson, Susanne Fronn˚a, Malin ˚Ashuvud, Moa ¨Onell and Jenny H¨agglund have all been very helpful to me. Per Nystr¨om and Tord Heljeberg, from the library of M¨alardalen University, have provided highly valued research support.
Last, but not least, I owe a lot of gratitude to my dear family and friends for all their positivity and love. The greatly skilled Naiara Seara created an excellent illustration for the front cover. My eminent wife Noelia Gutierrez Rodriguez together with Anders, Ingrid, Julia, Joel and Aileen Hellstr¨om have always believed in me and I have received immense support from everyone. This journey would never have been possible without you!
Per Anders Rickard Hellstr¨om V¨aster˚as, October 2019
vi
List of Publications
The following is a list of publications that
form the basis of this doctoral thesis
1Paper A (conference)
Intelligent Wireless Body Area Network System for Human Motion Analysis
Per Hellstrom, Lennie Carl´en Eriksson,
Jonatan Scharff Willners, Mia Folke, Martin Ekstr¨om
The First International Conference on Smart Portable, Wearable, Implantable and Disability-oriented Devices and Systems,
SPWID 2015, June 21-26, pages 37-40, 2015 Paper B (conference)
Wearable Weight Estimation System
Per Hellstrom, Mia Folke, Martin Ekstr¨om
The Fourth International Conference on Health and Social Care Information Systems and Technologies, HCIST 2015, October 7-9, Procedia Computer Science, volume 64, pages 146-152, 2015
1The included articles have been reformatted to comply with the dissertation layout
List of Publications
The following is a list of publications that
form the basis of this doctoral thesis
1Paper A (conference)
Intelligent Wireless Body Area Network System for Human Motion Analysis
Per Hellstrom, Lennie Carl´en Eriksson,
Jonatan Scharff Willners, Mia Folke, Martin Ekstr¨om
The First International Conference on Smart Portable, Wearable, Implantable and Disability-oriented Devices and Systems,
SPWID 2015, June 21-26, pages 37-40, 2015 Paper B (conference)
Wearable Weight Estimation System
Per Hellstrom, Mia Folke, Martin Ekstr¨om
The Fourth International Conference on Health and Social Care Information Systems and Technologies, HCIST 2015, October 7-9, Procedia Computer Science, volume 64, pages 146-152, 2015
1The included articles have been reformatted to comply with the dissertation layout
List of Publications
Paper C (conference)
Walking Intensity Estimation with a Portable Pedobarography System
Per Anders Rickard Hellstrom, Anna ˚Akerberg, Martin Ekstr¨om and Mia Folke
The Thirteenth International Conference on Wearable Micro and Nano Technologies for Personalized Health, pHealth 2016, May 29-31, Studies in health technology and informatics, volume 224, pages 27-32, 2016
Paper D (journal)
Evaluation of the IngVaL Pedobarography System for Monitoring of Walking Speed
Per Hellstr¨om, Anna ˚Akerberg, Martin Ekstr¨om and Mia Folke
Healthcare Informatics Research (HIR), pages 118-124, April 2018 Paper E (conference)
Carried Weight Affects Walking Speed Monitoring with the IngVaL System
Per Anders Rickard Hellstrom and Mia Folke
The Sixteenth International Conference on Wearable Micro and Nano Technologies for Personalized Health, pHealth 2019, June 10-12, Studies in health technology and informatics, volume 261, pages 317-320, 2019
Paper F (conference)
Monitoring of Carried Weight During Walk Using a Wearable Pedobarography System
Per Anders Rickard Hellstrom and Mia Folke
The Fifth International Conference on Smart Portable, Wearable, Implantable and Disability-oriented Devices and Systems, SPWID 2019, July 28-August 2, pages 5-8, 2019
Paper G (journal)
Novel Weight Estimation Analyses and the Development of the Wearable IngVaL System for Monitoring of
Health Related Walk Parameters
Per A. R. Hellstrom and Mia Folke
Submitted to Journal on Advances in Life Sciences, October 2019
viii
List of Publications
Paper C (conference)
Walking Intensity Estimation with a Portable Pedobarography System
Per Anders Rickard Hellstrom, Anna ˚Akerberg, Martin Ekstr¨om and Mia Folke
The Thirteenth International Conference on Wearable Micro and Nano Technologies for Personalized Health, pHealth 2016, May 29-31, Studies in health technology and informatics, volume 224, pages 27-32, 2016
Paper D (journal)
Evaluation of the IngVaL Pedobarography System for Monitoring of Walking Speed
Per Hellstr¨om, Anna ˚Akerberg, Martin Ekstr¨om and Mia Folke
Healthcare Informatics Research (HIR), pages 118-124, April 2018 Paper E (conference)
Carried Weight Affects Walking Speed Monitoring with the IngVaL System
Per Anders Rickard Hellstrom and Mia Folke
The Sixteenth International Conference on Wearable Micro and Nano Technologies for Personalized Health, pHealth 2019, June 10-12, Studies in health technology and informatics, volume 261, pages 317-320, 2019
Paper F (conference)
Monitoring of Carried Weight During Walk Using a Wearable Pedobarography System
Per Anders Rickard Hellstrom and Mia Folke
The Fifth International Conference on Smart Portable, Wearable, Implantable and Disability-oriented Devices and Systems, SPWID 2019, July 28-August 2, pages 5-8, 2019
Paper G (journal)
Novel Weight Estimation Analyses and the Development of the Wearable IngVaL System for Monitoring of
Health Related Walk Parameters
Per A. R. Hellstrom and Mia Folke
Submitted to Journal on Advances in Life Sciences, October 2019
viii
Contents
Acknowledgement v
List of Publications vii
I
Doctoral Thesis
1
1 Introduction 3 1.1 Motivation . . . 3 1.2 Aim . . . 4 1.3 Research Questions . . . 5 1.3.1 RQ1 (measurement system) . . . 5 1.3.2 RQ2 (carried weight) . . . 5 1.3.3 RQ3 (walking speed) . . . 5 1.4 Delimitations . . . 6 1.5 Research Methodology . . . 61.5.1 Ethical Experiment Reflections . . . 6
1.6 Research Contributions . . . 7 1.6.1 Paper A (RQ1) . . . 7 1.6.2 Paper B (RQ1 and RQ2) . . . 8 1.6.3 Paper C (RQ3) . . . 8 1.6.4 Paper D (RQ1 and RQ3) . . . 8 1.6.5 Paper E (RQ2 and RQ3) . . . 9 1.6.6 Paper F (RQ2) . . . 9 1.6.7 Paper G (RQ1 and RQ2) . . . 10 1.7 Thesis Structure . . . 10
Contents
Acknowledgement v List of Publications viiI
Doctoral Thesis
1
1 Introduction 3 1.1 Motivation . . . 3 1.2 Aim . . . 4 1.3 Research Questions . . . 5 1.3.1 RQ1 (measurement system) . . . 5 1.3.2 RQ2 (carried weight) . . . 5 1.3.3 RQ3 (walking speed) . . . 5 1.4 Delimitations . . . 6 1.5 Research Methodology . . . 61.5.1 Ethical Experiment Reflections . . . 6
1.6 Research Contributions . . . 7 1.6.1 Paper A (RQ1) . . . 7 1.6.2 Paper B (RQ1 and RQ2) . . . 8 1.6.3 Paper C (RQ3) . . . 8 1.6.4 Paper D (RQ1 and RQ3) . . . 8 1.6.5 Paper E (RQ2 and RQ3) . . . 9 1.6.6 Paper F (RQ2) . . . 9 1.6.7 Paper G (RQ1 and RQ2) . . . 10 1.7 Thesis Structure . . . 10
13
Contents
2 Background 13
2.1 Devices Used for Monitoring of
Physical Activity . . . 13 2.2 Pedobarography . . . 14 2.2.1 Foot Anatomy . . . 14 2.2.2 Gait . . . 15 2.2.3 Sensor Types . . . 16 2.3 Applications of Pedobarography . . . 16 2.4 Parameters of Interest . . . 17 2.4.1 Spatio-Temporal Parameters . . . 17
2.4.2 Force and Pressure Parameters . . . 18
2.5 Wearable Commercial PBG Systems Used in Research . . 19
3 Design of the Pedobarography System 21 3.1 System Requirements . . . 21
3.2 Design Choices . . . 22
3.2.1 Choosing Force Sensor Type . . . 22
3.2.2 Sensor Implementation . . . 23
3.2.3 Sensor Calibration . . . 25
3.2.4 Wireless Communication . . . 26
3.3 System Verification . . . 30
3.4 What is Possible for IngVaL to Measure? . . . 32
3.5 Summary of the System Design . . . 32
4 Validation of the Applications of the PBG System 33 4.1 Carried Weight . . . 33
4.2 Walking Speed . . . 38
5 Discussion and Conclusion 43 5.1 Discussion . . . 43
5.2 Conclusion . . . 46
Bibliography 49
II
Included Papers
61
6 Paper A: Intelligent Wireless Body Area Network System x Contents 2 Background 13 2.1 Devices Used for Monitoring of Physical Activity . . . 13 2.2 Pedobarography . . . 14 2.2.1 Foot Anatomy . . . 14 2.2.2 Gait . . . 15 2.2.3 Sensor Types . . . 16 2.3 Applications of Pedobarography . . . 16 2.4 Parameters of Interest . . . 17 2.4.1 Spatio-Temporal Parameters . . . 172.4.2 Force and Pressure Parameters . . . 18
2.5 Wearable Commercial PBG Systems Used in Research . . 19
3 Design of the Pedobarography System 21 3.1 System Requirements . . . 21
3.2 Design Choices . . . 22
3.2.1 Choosing Force Sensor Type . . . 22
3.2.2 Sensor Implementation . . . 23
3.2.3 Sensor Calibration . . . 25
3.2.4 Wireless Communication . . . 26
3.3 System Verification . . . 30
3.4 What is Possible for IngVaL to Measure? . . . 32
3.5 Summary of the System Design . . . 32
4 Validation of the Applications of the PBG System 33 4.1 Carried Weight . . . 33
4.2 Walking Speed . . . 38
5 Discussion and Conclusion 43 5.1 Discussion . . . 43
5.2 Conclusion . . . 46
Bibliography 49
II
Included Papers
61
6 Paper A:
Intelligent Wireless Body Area Network System
x
Contents
for Human Motion Analysis 63
6.1 Introduction . . . 64 6.2 System Description . . . 65 6.2.1 Hardware . . . 65 6.2.2 Data Communication . . . 70 6.2.3 Analysis . . . 72 6.2.4 Initial Measurements . . . 72 6.3 Discussion . . . 72 References . . . 73 7 Paper B: Wearable Weight Estimation System 77 7.1 Introduction . . . 79
7.2 Methods . . . 80
7.2.1 The Measurement System . . . 80
7.2.2 Experiment . . . 81 7.2.3 Analysis . . . 83 7.3 Results . . . 83 7.4 Discussion . . . 86 7.5 Conclusion . . . 87 References . . . 87 8 Paper C: Walking Intensity Estimation with a Portable Pedobarography System 91 8.1 Introduction . . . 92 8.2 Methods . . . 93 8.2.1 Hardware . . . 93 8.2.2 Experiment Setup . . . 94 8.2.3 Data Analysis . . . 94 8.3 Results . . . 95 8.4 Discussion . . . 97 8.5 Conclusion . . . 100 References . . . 100 9 Paper D: Evaluation of the IngVaL Pedobarography System for Monitoring of Walking Speed 103 9.1 Introduction . . . 105
Contents for Human Motion Analysis 63 6.1 Introduction . . . 64 6.2 System Description . . . 65 6.2.1 Hardware . . . 65 6.2.2 Data Communication . . . 70 6.2.3 Analysis . . . 72 6.2.4 Initial Measurements . . . 72 6.3 Discussion . . . 72 References . . . 73 7 Paper B: Wearable Weight Estimation System 77 7.1 Introduction . . . 79
7.2 Methods . . . 80
7.2.1 The Measurement System . . . 80
7.2.2 Experiment . . . 81 7.2.3 Analysis . . . 83 7.3 Results . . . 83 7.4 Discussion . . . 86 7.5 Conclusion . . . 87 References . . . 87 8 Paper C: Walking Intensity Estimation with a Portable Pedobarography System 91 8.1 Introduction . . . 92 8.2 Methods . . . 93 8.2.1 Hardware . . . 93 8.2.2 Experiment Setup . . . 94 8.2.3 Data Analysis . . . 94 8.3 Results . . . 95 8.4 Discussion . . . 97 8.5 Conclusion . . . 100 References . . . 100 9 Paper D: Evaluation of the IngVaL Pedobarography System for Monitoring of Walking Speed 103 9.1 Introduction . . . 105
Contents
9.2 Methods . . . 106
9.2.1 IngVaL Hardware Development . . . 106
9.2.2 Experiment . . . 109
9.2.3 Analysis . . . 110
9.3 Results . . . 112
9.4 Discussion . . . 115
10 Paper E: Carried Weight Affects Walking Speed Monitoring with the IngVaL System 121 10.1 Introduction . . . 122 10.2 Methods . . . 123 10.2.1 Hardware . . . 123 10.2.2 Experiment Setup . . . 123 10.2.3 Data Analysis . . . 125 10.3 Results . . . 125 10.4 Discussion . . . 126 10.5 Conclusion . . . 127 References . . . 127 11 Paper F: Monitoring of Carried Weight During Walk Using a Wearable Pedobarography System 129 11.1 Introduction . . . 130 11.2 Methods . . . 131 11.2.1 Hardware . . . 131 11.2.2 Experiment Setup . . . 134 11.2.3 Data Analysis . . . 135 11.3 Results . . . 136 11.4 Discussion . . . 136 11.5 Conclusion . . . 138 References . . . 138 12 Paper G: Novel Weight Estimation Analyses and the Development of the Wearable IngVaL System for Monitoring of Health Related Walk Parameters 141 12.1 Introduction . . . 142
12.2 System Development . . . 145
xii Contents 9.2 Methods . . . 106
9.2.1 IngVaL Hardware Development . . . 106
9.2.2 Experiment . . . 109
9.2.3 Analysis . . . 110
9.3 Results . . . 112
9.4 Discussion . . . 115
10 Paper E: Carried Weight Affects Walking Speed Monitoring with the IngVaL System 121 10.1 Introduction . . . 122 10.2 Methods . . . 123 10.2.1 Hardware . . . 123 10.2.2 Experiment Setup . . . 123 10.2.3 Data Analysis . . . 125 10.3 Results . . . 125 10.4 Discussion . . . 126 10.5 Conclusion . . . 127 References . . . 127 11 Paper F: Monitoring of Carried Weight During Walk Using a Wearable Pedobarography System 129 11.1 Introduction . . . 130 11.2 Methods . . . 131 11.2.1 Hardware . . . 131 11.2.2 Experiment Setup . . . 134 11.2.3 Data Analysis . . . 135 11.3 Results . . . 136 11.4 Discussion . . . 136 11.5 Conclusion . . . 138 References . . . 138 12 Paper G: Novel Weight Estimation Analyses and the Development of the Wearable IngVaL System for Monitoring of Health Related Walk Parameters 141 12.1 Introduction . . . 142
12.2 System Development . . . 145 xii
Contents
12.2.1 Selection of Sensor Type . . . 145
12.2.2 Sensor Implementation . . . 148
12.2.3 Calibration of the Sensors . . . 150
12.3 Method . . . 153
12.3.1 Experiment . . . 153
12.3.2 Data Analysis . . . 155
12.4 Results . . . 157
12.5 Discussion . . . 160
12.6 Conclusion and Future Work . . . 161
References . . . 162
Abbreviations 167 Contents 12.2.1 Selection of Sensor Type . . . 145
12.2.2 Sensor Implementation . . . 148
12.2.3 Calibration of the Sensors . . . 150
12.3 Method . . . 153
12.3.1 Experiment . . . 153
12.3.2 Data Analysis . . . 155
12.4 Results . . . 157
12.5 Discussion . . . 160
12.6 Conclusion and Future Work . . . 161
References . . . 162
Abbreviations 167
I
Doctoral Thesis
I
Doctoral Thesis
1
Introduction
Monitoring of health increases, at home and at work, to improve quality of life and to prevent unhealthiness. The proportion of the older adults in the population is increasing and thus also the healthcare costs [1]. The hospitals and the healthcare personnel will struggle to keep up with the increasing need for healthcare if we don’t increase the health monitoring at home and at work. This means a shift is needed towards preventative Personal Health Monitoring (PHM), outside of the hospital setting [2].
1.1
Motivation
Health is defined, by the World Health Organization (WHO), as a state of complete physical, mental and social well-being and not merely the absence of disease or infirmity. The main motivation for the research in this thesis is that PHM, at least the physical part, can be done with the help of sensors to give an idea of risks related to health. An important subcategory within PHM is human motion analysis, which Pedobarography (PBG) is a part of. Force sensors built into shoes, to monitor the forces between the feet and insoles, can be used for many things related to health. PBG started back in the late 19th century but the field is not perfected yet and there are many scientific challenges left to solve.
Two major health problems are sedentary behaviour and back problems. Sedentary behavior and physical inactivity cause an increased risk of many different diseases, especially the cardiovascular diseases [3]. Sedentary
1
Introduction
Monitoring of health increases, at home and at work, to improve quality of life and to prevent unhealthiness. The proportion of the older adults in the population is increasing and thus also the healthcare costs [1]. The hospitals and the healthcare personnel will struggle to keep up with the increasing need for healthcare if we don’t increase the health monitoring at home and at work. This means a shift is needed towards preventative Personal Health Monitoring (PHM), outside of the hospital setting [2].
1.1
Motivation
Health is defined, by the World Health Organization (WHO), as a state of complete physical, mental and social well-being and not merely the absence of disease or infirmity. The main motivation for the research in this thesis is that PHM, at least the physical part, can be done with the help of sensors to give an idea of risks related to health. An important subcategory within PHM is human motion analysis, which Pedobarography (PBG) is a part of. Force sensors built into shoes, to monitor the forces between the feet and insoles, can be used for many things related to health. PBG started back in the late 19th century but the field is not perfected yet and there are many scientific challenges left to solve.
Two major health problems are sedentary behaviour and back problems. Sedentary behavior and physical inactivity cause an increased risk of many different diseases, especially the cardiovascular diseases [3]. Sedentary
1. Introduction
behavior can also cause back pain [4]. Heavy work conditions, on the other hand, can damage the body and thereby cause back problems. The total amount of lifted weights and lift frequency are moderate to strong risk factors for lower back pain [5]. Pain in the lower back is one of the most common health problems today and is expected to become even more frequent in the future [6]. About a third of all employees in Sweden, during the year 2015, experienced pain in their lower back every week [7]. The same year, 16% of the employed men and 10% of the employed women lifted more than 15 kg several times a day [7]. Health problems could be decreased by long term monitoring and giving early feedback. Interesting walking parameters to measure are thus, among others, the walking speed and the amount of carried weight. PBG is an alternative that can be used for measurement of walking speed (and thus the step length), carried weight while walking and long term monitoring of the walking style to early discover health related problems. Decreasing walking speed and shorter steps, when monitoring over time, can give an early indication of a decreased health status in older adults [8]. Shuffling of the feet while walking and having a worse balance occurs more often for the older adults. Monitoring of balance is another important application since it enables early warning for an increased risk of falling [9]. Walking distance and carried weight are both important for estimation of how much energy that is spent while moving. The Energy Expenditure (EE) is the total amount of energy that leaves the body. EE is also affected by many other factors, such as the metabolic rate at rest and the fitness level, and more factors will be discussed in Section 3.4.
1.2
Aim
The aim of this thesis is: ”To design, build and evaluate a wireless
wearable measurement system based on pedobarography for
monitoring of health related walk parameters”.
A wireless data transfer enables the use of smartphone or tablet applic-ations. Long term monitoring of walking requires a wearable system. Pedobarography is the chosen technology since it is a common choice for evaluation of human walking motion. The two chosen applications for validating the system are monitoring of the carried weight during walk and monitoring of the walking speed.
4
1. Introduction
behavior can also cause back pain [4]. Heavy work conditions, on the other hand, can damage the body and thereby cause back problems. The total amount of lifted weights and lift frequency are moderate to strong risk factors for lower back pain [5]. Pain in the lower back is one of the most common health problems today and is expected to become even more frequent in the future [6]. About a third of all employees in Sweden, during the year 2015, experienced pain in their lower back every week [7]. The same year, 16% of the employed men and 10% of the employed women lifted more than 15 kg several times a day [7]. Health problems could be decreased by long term monitoring and giving early feedback. Interesting walking parameters to measure are thus, among others, the walking speed and the amount of carried weight. PBG is an alternative that can be used for measurement of walking speed (and thus the step length), carried weight while walking and long term monitoring of the walking style to early discover health related problems. Decreasing walking speed and shorter steps, when monitoring over time, can give an early indication of a decreased health status in older adults [8]. Shuffling of the feet while walking and having a worse balance occurs more often for the older adults. Monitoring of balance is another important application since it enables early warning for an increased risk of falling [9]. Walking distance and carried weight are both important for estimation of how much energy that is spent while moving. The Energy Expenditure (EE) is the total amount of energy that leaves the body. EE is also affected by many other factors, such as the metabolic rate at rest and the fitness level, and more factors will be discussed in Section 3.4.
1.2
Aim
The aim of this thesis is: ”To design, build and evaluate a wireless
wearable measurement system based on pedobarography for
monitoring of health related walk parameters”.
A wireless data transfer enables the use of smartphone or tablet applic-ations. Long term monitoring of walking requires a wearable system. Pedobarography is the chosen technology since it is a common choice for evaluation of human walking motion. The two chosen applications for validating the system are monitoring of the carried weight during walk and monitoring of the walking speed.
4
1.3 Research Questions
1.3
Research Questions
The first Research Question (RQ) is related to engineering and the two other RQs are related to the analysis of health related parameters during walk.
1.3.1
RQ1 (measurement system)
How can a pedobarography system for personal health monit-oring be designed and how can the durability be increased for force sensing resistors implemented in insoles?
Measurement systems for personal health should be cheap enough to be used at a larger scale. One way of reducing the cost of the system is by using a less amount of sensors. It is also important to implement the sensors in the insoles with the durability of the sensors in mind. The insole is the replaceable interior bottom of the shoe. The sensors should be kept in working order for as many steps as possible.
1.3.2
RQ2 (carried weight)
How should the analysis be done with the pedobarography sys-tem to monitor carried weight during walk?
The use of as few sensors as possible to reduce cost has the side effect of making it more difficult to estimate how much weight that is carried. The carried weight is monitored by measuring the total weight and then subtracting the body weight. Feedback to the worker will make them aware of how heavy the working conditions are.
1.3.3
RQ3 (walking speed)
How should the analysis be done with the pedobarography sys-tem to monitor walking speed?
Long term monitoring of changes in preferred walking speed, and step length, are interesting for early detection of health status changes. The time between two steps is easy to measure and the step frequency is the inverse of the step time. Step length is related to the force-time integral (impulse) measured during the last part of the contact with the ground (at toe-off) during a step.
1.3 Research Questions
1.3
Research Questions
The first Research Question (RQ) is related to engineering and the two other RQs are related to the analysis of health related parameters during walk.
1.3.1
RQ1 (measurement system)
How can a pedobarography system for personal health monit-oring be designed and how can the durability be increased for force sensing resistors implemented in insoles?
Measurement systems for personal health should be cheap enough to be used at a larger scale. One way of reducing the cost of the system is by using a less amount of sensors. It is also important to implement the sensors in the insoles with the durability of the sensors in mind. The insole is the replaceable interior bottom of the shoe. The sensors should be kept in working order for as many steps as possible.
1.3.2
RQ2 (carried weight)
How should the analysis be done with the pedobarography sys-tem to monitor carried weight during walk?
The use of as few sensors as possible to reduce cost has the side effect of making it more difficult to estimate how much weight that is carried. The carried weight is monitored by measuring the total weight and then subtracting the body weight. Feedback to the worker will make them aware of how heavy the working conditions are.
1.3.3
RQ3 (walking speed)
How should the analysis be done with the pedobarography sys-tem to monitor walking speed?
Long term monitoring of changes in preferred walking speed, and step length, are interesting for early detection of health status changes. The time between two steps is easy to measure and the step frequency is the inverse of the step time. Step length is related to the force-time integral (impulse) measured during the last part of the contact with the ground (at toe-off) during a step.
1. Introduction
1.4
Delimitations
The measurement system should be portable and use wireless transmis-sion of the collected data to enable the use of mobile phone or tablet applications. To make it possible to process the data on an embedded system, the proposed analyses methods should preferably be kept from becoming too complex. A cost effective measurement system is preferred for use in PHM. This means that the number of sensors should be kept low as it is one of the main factors for the total cost of the system. The sensors should be embedded in the insole, instead of put in a custom built solution inside a specific pair of shoes. Insoles can not be used in all types of footwear as the insoles are required to be held in place by the shoe. Clogs and sandals without a heel cap are examples of unsuitable footwear, since the insole can slip out.
1.5
Research Methodology
The first step was a literature study to examine the related work. The study enabled formulation of unsolved problems and research questions. The new research was built on previously known scientific results to avoid repeating work that had already been done. The research questions changed over time since they became more and more refined and addi-tional literature studies were made to address these changes. The research area is multidisciplinary and encompasses biomedical measurements, elec-tronics and computer science. Challenges appeared, while searching for answers to the identified research questions, and the work on solving the challenges resulted in research results. An iterative approach was used where the first prototype design of the measurement system (RQ1) was tested for the two chosen applications (RQ2 and RQ3) to validate the system. A deductive method was used where a failed validation of the system design meant that the system had to be improved or rejected. Flaws in the system design were addressed and the second system version was then used from Paper D and onwards. The second system version is called Identifying Velocity and Load (IngVaL).
1.5.1
Ethical Experiment Reflections
The collected data has to be stored for 10 years before being removed. A participant in a study may at any time, and without stating a reason, 6
1. Introduction
1.4
Delimitations
The measurement system should be portable and use wireless transmis-sion of the collected data to enable the use of mobile phone or tablet applications. To make it possible to process the data on an embedded system, the proposed analyses methods should preferably be kept from becoming too complex. A cost effective measurement system is preferred for use in PHM. This means that the number of sensors should be kept low as it is one of the main factors for the total cost of the system. The sensors should be embedded in the insole, instead of put in a custom built solution inside a specific pair of shoes. Insoles can not be used in all types of footwear as the insoles are required to be held in place by the shoe. Clogs and sandals without a heel cap are examples of unsuitable footwear, since the insole can slip out.
1.5
Research Methodology
The first step was a literature study to examine the related work. The study enabled formulation of unsolved problems and research questions. The new research was built on previously known scientific results to avoid repeating work that had already been done. The research questions changed over time since they became more and more refined and addi-tional literature studies were made to address these changes. The research area is multidisciplinary and encompasses biomedical measurements, elec-tronics and computer science. Challenges appeared, while searching for answers to the identified research questions, and the work on solving the challenges resulted in research results. An iterative approach was used where the first prototype design of the measurement system (RQ1) was tested for the two chosen applications (RQ2 and RQ3) to validate the system. A deductive method was used where a failed validation of the system design meant that the system had to be improved or rejected. Flaws in the system design were addressed and the second system version was then used from Paper D and onwards. The second system version is called Identifying Velocity and Load (IngVaL).
1.5.1
Ethical Experiment Reflections
The collected data has to be stored for 10 years before being removed. A participant in a study may at any time, and without stating a reason, 6
1.6 Research Contributions
get a copy of the data or have the link removed between the data and the participant. The data can not be traced to individual participants, without this physical linking document, which is stored in a secure location. Assessments were made regarding if the experiments would affect the volunteers physically and/or mentally. There was no risk for getting hurt mentally since only the performance of the system was analysed, and not the performance of the participants. Walking and carrying weights were considered to be safe but an ethical approval was requested and approved to make sure the risk analysis was correctly done. The ethical approval was granted by the Swedish Ethical Review Authority (diary number 2017/070). Every effort was made to conduct the experiments in a safe way. Great care was taken to protect the spine from the added weight inside the backpack. Two mobile phones were available during the experiments to ensure swift contact with healthcare services if an accident would have happened.
1.6
Research Contributions
This section presents the results of the included papers in this thesis and the contributions made by the main author of the papers, Per Hellstr¨om.
1.6.1
Paper A (RQ1)
Wireless body area network system with each node including an Inertial Measurement Unit (IMU) and the possibility to connect external sensors. IMU devices will be explained in Section 2.1. The main system was designed by the second and third authors of Paper A during their Master’s thesis work. A body area network is a wireless area network that is worn on the body.
Contribution by Per Hellstr¨om for Paper A
Per was responsible for the writing of the paper and contributed to the part of the electronics related to the add-on card for external force sensors. He presented the paper at the 1st International Conference on Smart Portable, Wearable, Implantable and Disability-oriented Devices and Systems (SPWID 2015).
1.6 Research Contributions
get a copy of the data or have the link removed between the data and the participant. The data can not be traced to individual participants, without this physical linking document, which is stored in a secure location. Assessments were made regarding if the experiments would affect the volunteers physically and/or mentally. There was no risk for getting hurt mentally since only the performance of the system was analysed, and not the performance of the participants. Walking and carrying weights were considered to be safe but an ethical approval was requested and approved to make sure the risk analysis was correctly done. The ethical approval was granted by the Swedish Ethical Review Authority (diary number 2017/070). Every effort was made to conduct the experiments in a safe way. Great care was taken to protect the spine from the added weight inside the backpack. Two mobile phones were available during the experiments to ensure swift contact with healthcare services if an accident would have happened.
1.6
Research Contributions
This section presents the results of the included papers in this thesis and the contributions made by the main author of the papers, Per Hellstr¨om.
1.6.1
Paper A (RQ1)
Wireless body area network system with each node including an Inertial Measurement Unit (IMU) and the possibility to connect external sensors. IMU devices will be explained in Section 2.1. The main system was designed by the second and third authors of Paper A during their Master’s thesis work. A body area network is a wireless area network that is worn on the body.
Contribution by Per Hellstr¨om for Paper A
Per was responsible for the writing of the paper and contributed to the part of the electronics related to the add-on card for external force sensors. He presented the paper at the 1st International Conference on Smart Portable, Wearable, Implantable and Disability-oriented Devices and Systems (SPWID 2015).
1. Introduction
1.6.2
Paper B (RQ1 and RQ2)
Design and implementation of a PBG measurement system based on commercial off the shelf components. Novel method for selecting meas-urement samples for weight estimation of carried load during walk. Ten participants did three walks, where two were with extra added weights and one without.
Contribution by Per Hellstr¨om for Paper B
Per was the main contributor and first author. He made part of the planning and was responsible for the measurements and analysis. Per designed and assembled the system. He presented the paper at the 4th International Conference on Health and Social Care Information Systems and Technologies (HCIST 2015).
1.6.3
Paper C (RQ3)
This paper was a pilot study for Paper D. The same system as in Paper B was used. A novel method was presented for relating the walking speed to the impulse, at the toe-off phase of the step, together with the step frequency. The toe-off happens at the last part of the contact with the ground and follows the heel strike and roll over phases of a step. This study included 10 participants and three walking speeds.
Contribution by Per Hellstr¨om for Paper C
Per was the main contributor and first author. He planned and was responsible for the measurements and the data analysis. He presented the paper at the 13th International Conference on Wearable Micro and Nano Technologies for Personalized Health (pHealth 2016).
1.6.4
Paper D (RQ1 and RQ3)
This study included 40 participants and five walking speeds to improve the certainty of the results in Paper C. A new type of reference system for the walking speed was used and a more comfortable version of insoles was designed. Milling away material in the insole, under the fragile boundary of the active sensor area, removed the possibility for putting mechanical pressure on this boundary. This resulted in an increased durability of 8
1. Introduction
1.6.2
Paper B (RQ1 and RQ2)
Design and implementation of a PBG measurement system based on commercial off the shelf components. Novel method for selecting meas-urement samples for weight estimation of carried load during walk. Ten participants did three walks, where two were with extra added weights and one without.
Contribution by Per Hellstr¨om for Paper B
Per was the main contributor and first author. He made part of the planning and was responsible for the measurements and analysis. Per designed and assembled the system. He presented the paper at the 4th International Conference on Health and Social Care Information Systems and Technologies (HCIST 2015).
1.6.3
Paper C (RQ3)
This paper was a pilot study for Paper D. The same system as in Paper B was used. A novel method was presented for relating the walking speed to the impulse, at the toe-off phase of the step, together with the step frequency. The toe-off happens at the last part of the contact with the ground and follows the heel strike and roll over phases of a step. This study included 10 participants and three walking speeds.
Contribution by Per Hellstr¨om for Paper C
Per was the main contributor and first author. He planned and was responsible for the measurements and the data analysis. He presented the paper at the 13th International Conference on Wearable Micro and Nano Technologies for Personalized Health (pHealth 2016).
1.6.4
Paper D (RQ1 and RQ3)
This study included 40 participants and five walking speeds to improve the certainty of the results in Paper C. A new type of reference system for the walking speed was used and a more comfortable version of insoles was designed. Milling away material in the insole, under the fragile boundary of the active sensor area, removed the possibility for putting mechanical pressure on this boundary. This resulted in an increased durability of 8
1.6 Research Contributions
the sensors and removed the chance of exaggerated values. The data acquisition system was changed from Android to Windows and the data transmission was improved.
Contribution by Per Hellstr¨om for Paper D
Per was the main contributor and first author. He planned and was responsible for the measurements and analysis in this study. Per also designed and assembled the new insoles and reference system. He came up with the new implementation for increased sensor durability. Per did smaller parts of the work regarding the conversion of the software from Android to Windows and regarding the automation of the calculations of the impulse integrals using a script written in the R language. The paper was published by the journal Healthcare Informatics Research in April 2018.
1.6.5
Paper E (RQ2 and RQ3)
Fifteen participants walked five one minute walks on a treadmill at a constant walking speed of 1.0 m/s. They carried five different added weights in a backpack. It is shown that the calculation used for predicting walking speed in Paper D also is affected by the carried weight. The relationship is linear as long as the participant is not fatigued.
Contribution by Per Hellstr¨om for Paper E
Per was the main contributor and first author. He planned and was responsible for the measurements and analysis. He presented the pa-per at the 16th International Conference on Wearable Micro and Nano Technologies for Personalized Health (pHealth 2019).
1.6.6
Paper F (RQ2)
This paper resulted an improved estimation of the weight during walk, compared to in Paper B. This improvement shows that the insoles that were introduced in Paper D, together with a improved calibration of the sensors, resulted in a better measurement performance. Fifteen participants walked five walks with a different total weight for each walk.
1.6 Research Contributions
the sensors and removed the chance of exaggerated values. The data acquisition system was changed from Android to Windows and the data transmission was improved.
Contribution by Per Hellstr¨om for Paper D
Per was the main contributor and first author. He planned and was responsible for the measurements and analysis in this study. Per also designed and assembled the new insoles and reference system. He came up with the new implementation for increased sensor durability. Per did smaller parts of the work regarding the conversion of the software from Android to Windows and regarding the automation of the calculations of the impulse integrals using a script written in the R language. The paper was published by the journal Healthcare Informatics Research in April 2018.
1.6.5
Paper E (RQ2 and RQ3)
Fifteen participants walked five one minute walks on a treadmill at a constant walking speed of 1.0 m/s. They carried five different added weights in a backpack. It is shown that the calculation used for predicting walking speed in Paper D also is affected by the carried weight. The relationship is linear as long as the participant is not fatigued.
Contribution by Per Hellstr¨om for Paper E
Per was the main contributor and first author. He planned and was responsible for the measurements and analysis. He presented the pa-per at the 16th International Conference on Wearable Micro and Nano Technologies for Personalized Health (pHealth 2019).
1.6.6
Paper F (RQ2)
This paper resulted an improved estimation of the weight during walk, compared to in Paper B. This improvement shows that the insoles that were introduced in Paper D, together with a improved calibration of the sensors, resulted in a better measurement performance. Fifteen participants walked five walks with a different total weight for each walk.
1. Introduction
Contribution by Per Hellstr¨om for Paper F
Per was the main contributor and first author. He planned and was responsible for the measurements and the analysis. He presented the paper at the 5th International Conference on Smart Portable, Wearable, Implantable and Disability-oriented Devices and Systems (SPWID 2019).
1.6.7
Paper G (RQ1 and RQ2)
This journal paper presented three novel analysis methods, and applied them to the same data as in Paper F, for measurement of weight during walk. The design process of the system was explained in detail and the design choices were motivated. Fifteen participants walked five walks with a different total weight for each walk.
Contribution by Per Hellstr¨om for Paper G
Per was the main contributor and first author. He was responsible for the data analysis. The paper was submitted to the Journal on Advances in Social Sciences in October 2019.
1.7
Thesis Structure
The structure of this doctoral thesis is as follows.
Part I of the Thesis
Part I of the thesis summarizes the research work and can be seen as an introduction to Part II, which contains the included papers.
Chapter 1 - Introduction
The first chapter presented the motivation for the work, the research aim and the three research questions, the delimitations, the general research methodology and the contributions of the author in the included papers. 10
1. Introduction
Contribution by Per Hellstr¨om for Paper F
Per was the main contributor and first author. He planned and was responsible for the measurements and the analysis. He presented the paper at the 5th International Conference on Smart Portable, Wearable, Implantable and Disability-oriented Devices and Systems (SPWID 2019).
1.6.7
Paper G (RQ1 and RQ2)
This journal paper presented three novel analysis methods, and applied them to the same data as in Paper F, for measurement of weight during walk. The design process of the system was explained in detail and the design choices were motivated. Fifteen participants walked five walks with a different total weight for each walk.
Contribution by Per Hellstr¨om for Paper G
Per was the main contributor and first author. He was responsible for the data analysis. The paper was submitted to the Journal on Advances in Social Sciences in October 2019.
1.7
Thesis Structure
The structure of this doctoral thesis is as follows.
Part I of the Thesis
Part I of the thesis summarizes the research work and can be seen as an introduction to Part II, which contains the included papers.
Chapter 1 - Introduction
The first chapter presented the motivation for the work, the research aim and the three research questions, the delimitations, the general research methodology and the contributions of the author in the included papers. 10
![Figure 4.1: Paper B: Estimation error [kg] on y-axis and actual weight [kg] on x-axis](https://thumb-eu.123doks.com/thumbv2/5dokorg/4747682.126065/53.718.125.589.315.689/figure-paper-estimation-error-axis-actual-weight-axis.webp)
