Bachelor’s Thesis

Interactive glove for mobility

training and rehabilitation
after stroke
Mihkel Hiob
Certec, Division of Rehabilitation Engineering Research
Department of Design Sciences
Faculty of Engineering LTH • Lund University • 2016
Interactive glove for mobility training and
rehabilitation after stroke

Mihkel Hiob

Certec, Division of Rehabilitation Engineering Research • Department of Design Sciences

Faculty of Engineering LTH • Lund University • 2016
Certec, Division of Rehabilitation Engineering research
Department of Design Sciences
Faculty of Engineering LTH, Lund University
P.O. Box 118
SE-221 00 Lund
Sweden

Certec report 1/2016

Preface

This bachelor’s thesis contains work completed between March and June 2016 at
Certec, Lund University, Sweden.
Several people have helped to contribute to the project.
I would like to thank my supervisor, Héctor Caltenco, for his valuable ideas, support
and feedback throughout the entire project.
Furthermore, I would like to thank everybody who agreed to come and test the
product. The feedback given was really helpful and has been taken into consideration
during the design process.

Lund, June 2016

Mihkel Hiob

i
Abstract

Stroke affects a vast amount of people every year. As a result, many survivors have at
least partially lost some of their abilities which are necessary to complete everyday
activities. Although rehabilitation can be used to improve those abilities, many
survivors claim that conventional methods are demotivating and discouraging.
The aim of this project is to develop and test interactive objects, which would
implement activities that are scientifically proven to have good effects for stroke
rehabilitation. Different exercises for stroke rehabilitation have been studied and their
feasibility of implementation evaluated. Furthermore, previous research and a few
existing solutions are also presented in this paper.
This thesis describes creating a glove using an iterative design process, which would
implement several hand exercises to control a computer’s mouse and keyboard. The
core of the glove is the Arduino Leonardo microcontroller, which is connected to
different sensors sewn into the glove. Hand exercises, such as wrist extension and
flexion or finger lifting, are used to control the computer in a simple way. For
example, to simulate an arrow key keystroke, the user has to apply pressure on his/her
fingertips where force sensors are located.
The final prototype was tested by multiple people. Although the user experience and
functionality of it could be improved, all participants said that they enjoyed using the
device. Hence, using the results from this project as well as from previous research, it
can be concluded that interactive objects can be motivational and encouraging for
mobility training and stroke rehabilitation.

Keywords: stroke, rehabilitation, interactive objects, Arduino, interactive design

iii
Table of Contents

1 Introduction ............................................................................................... 1
1.1 Strokes .................................................................................................................... 1
1.2 Stroke impacts and rehabilitation ............................................................................ 2
1.3 Motivation ................................................................................................................ 3
1.4 Research questions ................................................................................................ 3
1.5 Aim .......................................................................................................................... 4
2 Theoretical background ............................................................................ 5
2.1 Research on designing for stroke survivors ............................................................ 5
2.1.1 ActivABLES........................................................................................................................ 7
2.2 Existing solutions .................................................................................................... 8
2.3 Hand exercises for stroke rehabilitation ................................................................ 11
2.4 Tangible interaction ............................................................................................... 13
2.5 Vertical and horizontal prototyping ........................................................................ 14
3 Method...................................................................................................... 15
3.1 Research methods ................................................................................................ 15
3.2 Interaction design .................................................................................................. 15
4 Proposed solution and structured design ............................................. 19
4.1 Proposed solution ................................................................................................. 19
4.2 Structured design of functions .............................................................................. 20
4.3 Exercise implementation ....................................................................................... 21
5 Designing, prototyping and testing ........................................................ 23
5.1 Criteria and constraints ......................................................................................... 23
5.2 Materials ................................................................................................................ 23
5.2.1 littleBits............................................................................................................................. 23
5.2.2 Arduino microcontrollers .................................................................................................. 23
5.2.3 LilyPad accelerometer ..................................................................................................... 24
5.2.4 Sensors ............................................................................................................................ 25
v
5.2.5 Other materials ................................................................................................................ 27
5.3 Prototyping ............................................................................................................ 27
5.3.1 Low-fidelity prototype ....................................................................................................... 27
5.3.2 High-fidelity prototype ...................................................................................................... 29
5.3.2.1 Hardware implementation ............................................................................................. 31
5.3.2.2 Software implementation............................................................................................... 34
5.4 Testing................................................................................................................... 38
6 Conclusions ............................................................................................. 41
6.1 Discussion and conclusions .................................................................................. 41
6.2 Future work ........................................................................................................... 43
References .................................................................................................. 45
Appendix A : Schematic ............................................................................. 53
Appendix B : Code ..................................................................................... 55
Appendix C : Questionnaire....................................................................... 59

vi
1 Introduction

This chapter provides background information concerning strokes and the project.
Descriptions are provided regarding what strokes are, its impacts and rehabilitation.
The motivation and aim of the project are also introduced in this chapter.

1.1 Strokes
Stroke is a cerebrovascular disease which occurs when the blood supply to the brain
is cut off. In 2013, about 32.2 million people suffered from a stroke worldwide, 6.5
million of whom died [1].
The American Stroke Association [2], among other sources, identify a number of risk
factors for stroke:
 Age
Stroke usually affects older people, although there is a significant number of
younger people who also suffer from strokes. After the age of 55, the
probability of having a stroke doubles for each decade.
 Heredity
Stroke may be caused due to a gene mutation. Hence, the chances of having a
stroke are higher if a close relative has suffered from it.
 Race
Black people have the highest risk of death from a stroke as the probabilities
of having high blood pressure, diabetes and obesity are greater.
 Gender
Women are generally more likely to have a stroke than men.
 Other diseases
Artery and heart diseases, high blood pressure and cholesterol, diabetes and
many other heart conditions increase the chance of having a stroke.
 Bad habits
Smoking, alcohol and drug abuse, poor diet, physical inactivity and obesity
increase the risk of different diseases and stroke.
 Geographic location
Strokes are more prevalent in developing countries than in developed
countries [3]. While the main cause of strokes in developed countries is
ageing, the increase in prevalence in developing countries is due to smoking,
obesity and not being able to prevent a stroke due to lack of modern medical
devices [4].
 Socioeconomic factors
1
1 Introduction

Previous studies have shown that strokes are more common among lower
income people.
There are three types of strokes: ischaemic stroke, haemorrhagic stroke and transient
ischaemic attack (TIA) [5]. Firstly, an ischaemic stroke occurs when the blood supply
to the brain is cut off due to a blockage and accounts for about 87 percent of all cases
[6, 7]. The blockage can be caused by a blood clot developing at the clogged part of
the vessel, which is known as cerebral thrombosis. A blockage can also be caused by
a blood clot or other matter formed at another part of the circulatory system, which is
known as cerebral embolism [6]. Secondly, a haemorrhagic stroke results from a burst
of a blood vessel which causes bleeding in or around the brain and can be more
dangerous than an ischaemic stroke [8]. Thirdly, a TIA is caused by a temporary clot
[5]. Although the symptoms of a TIA may not last too long, it should be taken very
seriously as it may be a warning sign of a heart condition.

1.2 Stroke impacts and rehabilitation

As a result of a stroke, the nerve cells in the affected area are damaged, and without
oxygen, they start to fail and die after a few minutes [9]. This causes parts of the body
that the cells control to stop functioning, resulting in motor, sensory and cognitive
problems. The effects are often permanent because dead brain cells are not replaced
[10]. Hence, depending on where a stroke occurs and the extent of brain tissue
affected, one may have problems with moving the right side of the body, talking or
understanding speech (all controlled by the left side of the brain), vision or moving
the left side of the body (both controlled by the right side of the brain) [11]. When a
stroke occurs in the brainstem, it can affect eye movement, swallowing, breathing,
alertness and other specialised functions [11].
Due to the resilience and capabilities of the brain, it is possible to recover at least
some functionalities. Rehabilitation is an integral part of the recovery process and
begins at the hospital. It usually consists of physical, cognitive, emotional activities
and experimental therapies [12]. However, rehabilitation is different for each person,
as the type and severity of the stroke, general health, personality and other support
received, e.g. from family, need to be considered [13].
The goal of rehabilitation is to decrease the effects of a stroke in order to regain
independence and to attain the best potential quality of life, e.g. being able to do
everyday activities such as cooking, getting dressed and washing. Not all stroke
survivors need rehabilitation because of the low impact of a stroke, such as a TIA, but
those who do are expected to complete a number of activities over a period of time.
This involves many passive and active exercises. Passive exercises are those where a
patient is helped and assisted by a therapist, while active exercises are performed by
the patient independently [14]. In order to achieve the best outcome, both methods
should be combined. However, since rehabilitation might take months or years
depending on the severity and type of stroke, the repetitive nature of the exercises
may become frustrating. Furthermore, it is very difficult to stay motivated if progress
is not immediately evident. According to a study, only 31% of stroke survivors
perform home exercises prescribed by therapists due to a lack of motivation and slow

2
 1 Introduction

progress [15]. Hence, technology-assisted rehabilitation has become more widespread

in recent years to make the training easier, enjoyable, more motivating and
encouraging [16].
The most common physical impact of stroke is loss of limb movement and muscle
weakness [17]. For example, about 85% of people have problems with arm
movements after a stroke [18]. Hence, motion exercises are very common in order to
strengthen stroke impaired limbs [14].

1.3 Motivation
Stroke affects millions of people around the world every year. Those who survive
have generally at least some partial disability and hence may not be able to complete
some everyday activities such as cooking, cleaning, washing, etc. For example, a
study conducted by Wade and Hewer [19] revealed that 88% of people surveyed
relied on other people to perform everyday activities. This was determined by using
the Barthel index [20], a 20-point scale used to measure performance in activities of
daily living, one week after a stroke. Anything less than 20 points indicates
dependence on others. Hence, many people rely on rehabilitation in order to recover
at least some of their previous abilities. Since constant exercising is important, being
dedicated and motivated is vital in order to achieve the best result possible. Therefore,
rehabilitees should also perform active exercises at home. However, they may not be
confident enough and give up on rehabilitation which in turn may lead to incomplete
recovery. Additionally, a study conducted in the United States by Shaughnessy and
Resnick [21] found that 69% of stroke survivors do not perform all of the required or
prescribed exercising at home due to a lack of self-efficacy and motivation.
Furthermore, the repetitive nature of activities may also make them boring and
rehabilitees may lose interest in performing them. Thus, interactive objects that
motivate performing such exercises could be used to help in stroke rehabilitation.
Their flexibility can be used to implement multiple activities in an interesting and
motivating way, e.g. in order to change the TV channel, a user has to perform an
exercise beforehand. Therefore, it might help to keep the motivation up and in turn
result in better recovery. However, it should be kept in mind that not all stroke
survivors value the same things, and hence not all of them find the same exercises
motivating. Furthermore, by creating interactive objects which are easy to use, giving
clear instructions and feedback, survivors can be confident that they are performing
exercises correctly. Unfortunately, interactive objects are not very widely used for
stroke rehabilitation, but seem to be growing in popularity.

1.4 Research questions

How can interactive objects be successfully used to implement stroke rehabilitation
exercises?
How can interactive objects make performing repetitive tasks more fun and
motivating compared to performing them without any assistive objects?

3
1 Introduction

1.5 Aim
The aim of this project is to develop and test an interactive object which would
implement an activity that is already known to have good effect for stroke
rehabilitation. The interactive object should be robust, relatively compact, easy to
handle and should arouse one’s interest in using it. However, since many stroke
survivors find rehabilitation exercises rather boring and frustrating, one of the main
goals would also be that the interactive object would motivate them and make
performing exercises easier.
Age, physical and mental abilities as well as open-mindedness towards technology
vary a lot among stroke survivors. Hence, another goal of this project is to develop an
interactive object which could serve a vast amount of people.

4
2 Theoretical background

This chapter summarises some of the research on the field, as well as gives examples
of products already used for stroke rehabilitation. Some rehabilitation exercises are
also presented in this chapter. Furthermore, it gives an overview of tangible
interaction and how it is taken into account when designing interactive objects for
stroke rehabilitation. The concept of vertical and horizontal prototyping is introduced
at the end of the chapter.

2.1 Research on designing for stroke survivors

As some studies have shown, many stroke survivors are unmotivated or unconfident
to perform rehabilitation exercises at home. Hence, the amount of research on the
field is constantly increasing in order to find ways to make rehabilitation more
motivating by the use of technology. This usually involves stroke survivors early in
the design and testing processes to determine what they appreciate about a product
and what could be improved upon.
Axelrod et al. [22] carried out a project where they found that many elderly people
living with stroke could not see the benefits of technology, or that they were not
confident enough to use it due to over-complications. This is confirmed by a study by
Bjørkquist et al. [23] where elderly stroke survivors say that they do not feel entirely
comfortable with using telecare or telehealth technologies, but nevertheless would be
willing to use it if it would help them with stroke rehabilitation. As some
rehabilitation products might also need extra technology, e.g. a TV, some people were
also concerned about financial implications and extra space needed [22, 24]. Hence,
stroke rehabilitation products should not look very technical, and they should have
enough guidance such that stroke survivors would be comfortable and capable of
familiarising themselves with them. Furthermore, there should also be motivation as
well as space to perform physical exercises.
Curtis et al. [25] developed interactive games which would help stroke survivors with
handwriting rehabilitation. There were four games with adjustable levels of difficulty.
A user-centred design approach was used to target both stroke survivors and
healthcare specialists effectively. The application also provided a performance
history, audio-visual feedback, and posture reminders. Finally, the preliminary results
show that the games were both fun and motivating.
Balaam et al. [26] carried out a study where they worked closely with a few stroke
survivors and developed products which the participants thought they would find
interesting. In the beginning, the participants used their products actively, however,
5
2 Theoretical background

after a while some stopped using them as often because they claimed that doing the
same thing became boring. Some participants also reported that they did not like the
aesthetics of the products and thought that they would not fit in with their homes [24,
26]. Hence, it is important to have deep discussions with stroke survivors in order to
determine what they are interested in and involve them in designing and prototyping
stages so that the final product could be tailored according to their needs [23, 24, 26].
It is also important to consider other factors such as work and free time as well as
where and when the product will be used, and what other support is available, e.g.
family members [24, 26].
Delbressine et al. [27] designed a tangible table-top interaction with wearable
technology which would help rehabilitees with arm-hand rehabilitation by performing
grasping exercises. Many survivors reported that they both enjoyed using the object
and have found it to be useful as well. Furthermore, a study conducted by Michaelsen
et al. [28] shows that stroke rehabilitation might even be very beneficial for stroke
survivors with chronic hemiparesis.
Timmermans et al. [29, 30] have evaluated the feasibility and effects of technology-
supported task-orientated arm training (T-TOAT) and how it affects survivors’
motivation. The method comprises of separating training skills into functional
components such that the relationship with the original training skill is retained.
Figure 2.1 shows how skills such as drinking from a cup or eating with a knife and
fork are decomposed. The difficulty of each exercise is gradually increased based on
principles of exercise physiology and motor learning. Decomposing skills gives the
ability to implement exercise programs in technology-supported training. Participants
start with training on components individually, and after a while, the complete action
is trained. Timmermans et al. tested the method on nine chronic stroke survivors over
an eight-week training period. The results show that the method significantly
improved arm-hand performance as shown by mean values of Fugl–Meyer (14.2%
improvement, 𝑝 < 0.001), ARAT (15.3% improvement, 𝑝 < 0.05), MAL-AU
(43.4% improvement, 𝑝 < 0.05), and MAL-QU (34.1 % improvement, 𝑝 < 0.01),
and the results lasted after the training. Furthermore, the participants claimed that
their quality of life had improved, their motivation levels were higher and the system
was easy to use. Thus, the method is feasible and has positive effects on stroke
rehabilitation.

6
 2 Theoretical background

Figure 2.1 T-TOAT: decomposing skills [30]

Fitzpatrick et al. [24] mention that when designing an interactive object, it is also
important to consider the following:
 Balancing needs and wishes, e.g. rehabilitation versus what the patient likes
to do.
 Respecting the patient, i.e. what, how and when they want to do something.
 Determining what is fun and motivating for rehabilitees.
 Using a general design which would then be modified to specific needs.
 Creating a design which would be motivating in the long-term and flexible
enough to adapt to changing rehabilitation needs.
Therefore, stroke survivors are interested in exercises which would relate to their
favourite activities, e.g. reading books or playing chess, making them feel confident,
allowing them to be active and also involving other family members. All of this will
be taken into account during the designing stage.

2.1.1 ActivABLES
ActivABLES: Tangible Interaction to Support Effective and Usable Stroke
Rehabilitation at Home [31] is a project carried out at Certec, researching on how
physical and interactive objects can be used to make post-stroke rehabilitation easier
and more motivating. The project tries to develop tangibles which would implement
exercises already known to have a good effect on rehabilitation and help with
rehabilitation by incorporating them into stroke survivors’ daily lives [31]. The
project has four objectives [32]:
1. Investigating how different design methods can be used to create
rehabilitation tools which would be useful, enjoyable and motivating by
involving stroke survivors as well as other people such as caretakers in the
design process.
2. Developing interactive objects which would use multimodal feedback.

7
2 Theoretical background

3. Developing software which would incorporate interactive objects in daily

activities such as controlling the TV.
4. Investigating how interactive objects can be incorporated into stroke
survivors daily lives and their support networks in an effective way, e.g. to
keep rehabilitees motivated.
The project follows a user-centred design by conducting interviews and focus groups
with stroke healthcare professionals, stroke survivors and their families [33].
Interviews with stroke healthcare specialists have revealed that in order to get good
results, exercises should be meaningful, performed in context and part of an activity
[33]. Furthermore, it was found that reminders of performing an exercise are
important as well [33]. As already mentioned, the interviews have confirmed that
exercising should be fun, motivating, safe, individually tailored, and also balanced,
i.e. not too little but not too much either [33]. Furthermore, interactive objects should
also be empowering and encouraging, aesthetic, simple and easy to use [33].

2.2 Existing solutions

There are a number of solutions already developed for stroke rehabilitation. This
includes wearables which the user cannot interact with and is mainly used for
measurements, e.g. muscle strength. An alternative solution is interactive objects that
are used for exercising. Furthermore, some devices can be used for both
measurements and exercising, or just to help with specific things, such as walking.
Hence, interactive objects can be very encouraging whilst performing active
exercises. However, the biggest disadvantage amongst solutions currently is the lack
of specialised equipment [34].
Neofect in South Korea has developed a smart glove to help people with neurological
brain injuries and hand rehabilitation [35]. The wearable glove, shown in Figure 2.2,
comes with a software application that can evaluate a range of motions as well as
analyse motions of the fingers and hand [35]. Additionally, it provides the user with a
number of game-like exercises which implement various functional movements, such
as those related to daily tasks (wrist flexion/extension, forearm pronation/supination,
radial-ulnar deviation, and finger flexion/extension), in an effective and fun way [35,
36]. The software adjusts the level of difficulty of a game according to the learning
schedule algorithm which tries to balance between challenge and motivation [35].
The games database is updated monthly and consists of games like squeezing an
orange, pouring wine, catching balls or painting fences [36, 37]. Rehabilitees are also
able to see their performance results, such as their current state and improvements, in
a visual way [35].

8
 2 Theoretical background

Figure 2.2 RAPAEL Smart Glove™ [38]

The glove itself uses an Ultra-low-power ARM Cortex-M3 microcontroller, bending
sensors (variable resistors) and a 9-axis inertial measurement unit which measure
individual movements. The data is sent wirelessly to a computer via Bluetooth [35].
Furthermore, it is ergonomic and weighs only 132 grams [35].
Shin et al. conducted a study [36] where 46 stroke survivors were either using the
glove or conventional intervention for hand and arm rehabilitation. The study showed
that those using the glove improved significantly while the stroke survivors using the
conventional intervention method did not make any significant improvements. Hence,
it can be said that interactive objects have the potential in motivating rehabilitees as
well as being more effective.
Bioness is an American company which designs devices for people with stroke,
multiple sclerosis and other central nervous system disorders in order to regain
mobility and independence [39, 40, 41]. The company has developed three products
for stroke rehabilitation: the L300 for foot drop, the L300 Plus for foot drop plus
thigh weakness and the H200 for hand paralysis therapy [39].
The L300 foot drop system, shown in Figure 2.3, uses wireless technology that senses
how a person is walking and automatically adjusts to changes in terrain and walking
speed [42]. It stimulates the nerves and muscles to lift the foot in order to walk more
naturally [42]. The L300 Plus system adds a thigh cuff which helps to reduce thigh
weakness and increasing control over bending and straightening the knee [43]. The
H200 system, shown in Figure 2.4, uses wireless technology to help rehabilitees with
regaining hand function such as reaching, grasping, opening and closing the hand
[44]. Two studies by Hausdorf and Ring [45, 46] have confirmed that both L300
9
2 Theoretical background

systems significantly improve stability and walking speed compared to not using
them.

Figure 2.3 The L300 Foot Drop System [47]

Figure 2.4 The H200 for Hand Paralysis [48]

Saposnik et al. [49] used the Nintendo Wii gaming system to determine if it is
feasible, safe and efficacious for arm rehabilitation after a stroke compared to
recreational therapy. The system consists of a wireless controller and an infrared

10
 2 Theoretical background

sensor. While the user moves the controller with their arm, hand and wrist, the
sensors inside the controller measure speed, direction and acceleration, which are then
sent to the system. It is computer assisted and hence does not require big sweeping
movements. A TV screen provides real-time feedback, i.e. movements, and games are
therefore fun and motivating to perform. The study showed that using the Wii for
rehabilitation was more efficient than recreational rehabilitation as people were faster
and had stronger grip.

2.3 Hand exercises for stroke rehabilitation

There are many exercises which can help with stroke rehabilitation. However, since
the majority of stroke survivors suffer from problems with their hands, only those
related to that are presented. Furthermore, as will be discussed in the upcoming
chapters, this project deals with hand rehabilitation.
Figure 2.5 shows a wrist exercise. This involves lifting the wrist up, down and
sideways [50, 51]. Hence, the instructions for this exercise are quite straightforward
and easy. Furthermore, the other hand can be used to help with this exercise.

Figure 2.5 Wrist flexion and extension [50]

Another exercise is a wrist motion exercise. This involves turning the wrist in circles,
as shown in Figure 2.6, to encourage greater motion in the wrist [50]. Again, this
exercise does not require remembering many instructions.

11
2 Theoretical background

Figure 2.6 Wrist supination and pronation [50]

It is also important to strengthen the finger muscles. This can be achieved by
performing finger flexion and extension exercises. One of them is shown in Figure
2.7. The exercise consists of bending the fingers of the weaker hand into the palm and
then straightening them out again [51]. Another finger exercise is to lift fingers up
and down, as shown in Figure 2.8 [52].

Figure 2.7 Finger flexion and extension [51]

Figure 2.8 Finger lifting [52]

12
 2 Theoretical background

Many daily activities, such as holding heavy pans or opening jars, rely on grip
strength. Hence, the grip exercise shown in Figure 2.9 can be helpful to improve
flexibility and suppleness [53].

Figure 2.9 Grip strengthening [53]

2.4 Tangible interaction

Tangible interaction is an interdisciplinary field that studies user interfaces that are
physically embodied, i.e. when physical objects are used to interact with digital
information. For example, moving the computer mouse on a surface also moves the
pointer on the screen. Hence, tangible interaction is a great tool for stroke
rehabilitation as it provides instant feedback which both encourages and motivates
exercising. Thus, the following should be considered when designing tangibles for
stroke rehabilitees:
 Weight – physical object should be light so that the user does not feel
uncomfortable or become tired quickly.
 Size – smaller objects are generally easier to handle.
 Aesthetics – the object should be reasonable and fit in at home.
 Simplicity – instructions should be clear and easy to follow.
 Feedback – the user should be aware whether he or she is using the object in
a correct way.
Hornecker and Buur [54] say that tangible interaction share the following
characteristics:
 Tangibility and materiality – the user should be able to hold or touch the
object which uses senses, such as audio or visualisations, to give feedback.
 Physical embodiment of data – digital information is represented physically.

13
2 Theoretical background

 Embodied interaction – interaction experience should not only take into

account material settings but also social interaction and how it is embedded
in real space.
 Bodily movement – the user uses their own body to create movements and
gestures which would transfer digital data.
They also state that the following is important to consider when prototyping tangible
interaction:
 Sensuousness – how you input data, e.g. voice or movements, and when and
how you get feedback, e.g. instant visual feedback.
 Graduality – whether interaction is weak or strong, e.g. just having two
states, on and off, or having multiple input stages.
 Tightness – when and where do input and feedback occur, e.g. instant
feedback in the same room versus feedback later in another room.
 Behaviour and adaptation – whether the behaviour of the object is
predictable or confusing, and whether it is boring or makes the user angry.
Unfortunately, there are not many studies dealing with the use of tangible user
interfaces for stroke rehabilitation. Nevertheless, it will provide designers with many
opportunities and challenges. For example, tangible interaction gives the possibility to
create portable devices which users can carry along with them. However, more
research needs to be done on how to adjust tangible interaction for stroke survivors.

2.5 Vertical and horizontal prototyping

Vertical and horizontal prototyping can be used to implement interactive objects.
Horizontal prototyping provides a wide range of features which lack in detail. In
contrast, vertical prototyping shows fewer functions with more detail. These are
illustrated in Figure 2.10. Horizontal prototyping focuses on user interaction rather
than on functionalities, and are useful for understanding relationships and abilities of
the system. Furthermore, it helps to estimate development cost and time. Vertical
prototypes, on the other hand, are beneficial for exploring a complex function in
detail.

Figure 2.10 Vertical and horizontal prototyping [55]

14
3 Method

This chapter describes methods used in the project. It gives an overview of interaction
design and research methods used in this project

3.1 Research methods

This project uses mainly qualitative research methods, although quantitative research
methods are also used when necessary. As this paper draws a lot upon previous
research, existing literature is studied to gain an understanding of different activities
used for stroke rehabilitation and how interactive objects could be helpful. Previous
implementations of interactive objects are thoroughly studied and any relevant points
are used in the project.
Qualitative research methods are important to analyse human behaviour. Hence,
rather than quantifying data, it is essential to answer questions such as who, what,
when, where, why, how. This is essential when creating interactive objects for stroke
survivors and determining how they should be designed to be easy to use as well as
motivating. As it will be mentioned in Chapter 3.2, abled-bodied users are used to test
the prototypes and data is collected using questionnaires to gain an understanding of
how they perceive the interactive interface. Furthermore, literature on previous case
studies and interviews are also studied to gain an understanding of how to design
interactive objects for stroke rehabilitation.
Quantitative research expresses its data in numerical form. Since this project is
mainly concerned with how people perceive the interactive objects developed, e.g.
motivation, encouraging and fun, quantitative research is not that extensively used.
However, it will be useful when conducting background research and can be used to
support qualitative research.
This project relies mostly on empirical research as a lot of knowledge is obtained
from experiments and observation, i.e. users who have tested the prototypes.
However, exploratory research is also used to identify and define unclear questions.

3.2 Interaction design

The main focus of interaction design is on user experience. Cooper et al. [56, p.
xxviii] say that interaction design “is concerned most significantly with satisfying the
needs and desires of the people who will interact with a product or service.” Hence, a
product should not be confusing, ineffective or difficult to use, but rather the

15
3 Method

opposite. It should be enjoyable whilst helping the user in an effective and useful
way.
People’s strengths and weaknesses should be taken into account when designing a
product to ensure as good of a user experience as possible. It is also important to
consider who is going to use the product and how and where it is going to be used.
However, different people value different things, and hence, finding a design that
suits everybody is impossible. Thus, whilst designing, some compromises must be
made in order to find a good outcome.
Interaction design consists of four repeated activities [57, pp. 329-330]:
1. Establishing requirements – determining target users and their needs.
2. Designing alternatives – creating ideas that would meet the requirements.
3. Prototyping – creating samples of the alternatives.
4. Evaluating – determining how well the alternatives match the requirements.
An iterative design approach is used in this project as it is both time- and cost-
effective [58]. It consists of designing, prototyping and evaluation stages, as shown in
Figure 3.1.

Figure 3.1 Iterative design process

User action is taken into account when designing an interactive object. Though stroke
survivors will not be testing a prototype, previous knowledge obtained from user
studies of the ActivABLES project is used for designing the prototypes. Therefore,
even though it is not strictly a user-centred design process, the target users are always
kept in mind.
A prototype will be constantly upgraded to meet the design requirements and employ
results obtained at different stages of user testing. The fidelity of the prototype will
increase during the development process, starting with more crude Lo-Fi prototypes
gradually morphing into Hi-Fi prototypes. Low-fidelity prototypes are simple, take
little resources to create and provide high-level feedback. However, their function is
16
 3 Method

limited. As the project progresses, high-fidelity prototypes become more useful.

Although they require more time and money, they have lots of detail and
functionality. Since they are rather similar to the final product, they can be
successfully used for user testing.
Testing can start when the first prototype is complete. It provides important
information to go back upon whilst improving the design and prototype. However, it
should also be kept in mind that too much tester involvement may also cause
problems such that the final product becomes useless and ineffective for general
purposes [57, p. 324]. Therefore, at least two abled-bodied people who do not know
the details of the project will be involved in testing each prototype iteratively. During
each testing session, they will complete a pre-test and a post-test questionnaire. The
pre-test questionnaire is used to obtain background data of the participants. The post-
test questionnaire obtains qualitative information about the testing experience and
also uses the system usability scale [59]. The results will be taken into account to
improve designs and upgrade prototypes. Although it requires some time to conduct
the tests, using a testing-based iterative design process ensures that the final product
is more likely to be effective and easy to use.

17
4 Proposed solution and structured design

This chapter describes the idea proposed for implementation as well as its
advantages and disadvantages. Additionally, it gives an overview how structured
design is used in this project.

4.1 Proposed solution

This project focuses on implementing a glove that would help stroke survivors with
hand rehabilitation.
The idea is that a user would wear a glove with different sensors, such as flex and
force sensors as well as an accelerometer. When the user moves their hand, fingers or
squeezes something, the force applied could be determined using the sensors. Hence,
it is a good way to give the user direct feedback which is an important part of
rehabilitation.
This solution was chosen as around 85% of stroke survivors have loss of arm
movement, many of whom claim that rehabilitation does not pay enough attention to
arm rehabilitation [60]. Since arms are important for completing daily activities, arm-
hand rehabilitation exercises and recovery of functions can be beneficial and is
associated with improved quality of life [30, 60]. Hence, the glove could be used by a
large range of people. Furthermore, the glove itself would not be too technical or
intimidating, and should be suitable for non-technical people. Also, it would both be
cognitively and physically easy to handle and suitable for use by stroke survivors.
The glove would not weigh too much and is relatively small in size and would
therefore be comfortable to wear. The wires connected to the final product would be
hidden so it will not be aesthetically unpleasing, and can be stored in a drawer or
cupboard if needed.
The glove itself offers many opportunities. It can be implemented to do many tasks as
it can be used for both rehabilitation as well as for controlling other devices, such as
turning the TV on or off. Hence, this interactive object could potentially implement
more functions than the glove discussed in Chapter 2.2. Additionally, the latter is
rather expensive and mainly directed at healthcare institutions.
Thus, implementing a glove has many advantages. In addition to being easy to
handle, it is also rather flexible and can be used to perform many exercises as well as
help with different tasks. The ease of use is also a contributing factor that can
encourage and motivate users.

19
4 Proposed solution and structured design

4.2 Structured design of functions

In order to organise work and make finding errors as easy as possible, vertical and
horizontal prototyping and structured design methods are used.
Horizontal prototyping is used to get an overview of the project’s concept, while
vertical prototyping is used to conduct the project in order to create a product which
would be straightforward to use and would not confuse people with its wide range of
functionalities. Furthermore, it is likely that having only one functionality would help
to conduct user tests in a clear and easy way such that the results could be applied to
designs without overcomplicating them.
Structured design gives an overview and helps to organise the project as it breaks the
project down into several functions, showing the details inside them. In addition to
making planning and working more efficient, it also helps to find errors as well as to
distinguish important and unimportant parts.
Figure 4.1 shows a structured design for the vertical and horizontal prototyping of the
project. Initially, three proposals were considered.
Firstly, the home automation function would involve using the glove to control
devices at home. For example, by turning a hand either left of right, one can make the
light in the room become either brighter or dimmer. This solution would mean that
the users are encouraged to use the glove on a daily basis in order to achieve their
goals, e.g. to turn the lights on and off. However, there are several shortcomings.
Firstly, it might be difficult to integrate the glove with devices at home as each one
comes with different functions which need to be considered. Secondly, performing
exercises would be heavily dependent on the device and user’s lifestyle. This means
that the glove would be only be beneficial when it controls a device which the user
frequently needs.
Secondly, the music playing function would use hand gestures in conjunction with
computer software to create simple music. This function would be probably both
encouraging and motivating. Being able to perform fun activities is very useful for
rehabilitation and in turn helps people to achieve a better quality of life. However, as
the glove needs to be easy to use, it cannot use too many or too complicated hand
gestures. Hence, not a large number of different music sounds would be available.
This in turn might become boring after some time, especially for those with less
interest in music. Alternatively, those with previous musical talent may become
frustrated as the glove would not enable them to play to the ability that they have
been used to. Furthermore, some additional processing would be necessary to
compensate against sounds which may not be pleasing to human ears.
Thirdly, the game-playing function would use the glove to control computer inputs,
such as a mouse and keyboard, to play simple online games. There are a variety of
games available for different age groups and tastes. Hence, it is likely that the user
does not get bored as quickly as he or she has the ability to change a game whenever
needed. As the number of computer instructions, such as a mouse click or pressing an
arrow key is not vast, the product is likely to be easy to use. Additionally, it is a good
way to encourage people to repeat exercises multiple times. However, using
20
 4 Proposed solution and structured design

computers and playing online games might be intimidating for older people.
Furthermore, each game has different instructions and may thus be difficult to learn.
This function also assumes having a computer at home, which is not that uncommon
anymore.
Considering the criteria and constraints listed in this and previous chapters, the home
automation and music playing functions were discarded as they were too difficult to
implement in the project’s timeframe, as shown in Figure 4.1 (dashed lines). The
game-playing function (solid line) is implemented as it can be well used for
implementing the exercises outlined in Chapter 2.3.

Home Turn on home

automation appliances

Produce
Playing music different
sounds
Glove
Move the
cursor

Play simple
Imitate a
Playing games computer
mouse click
games

Imitate arrow
key keystrokes

Figure 4.1 Structured design for the prototype (the dashed paths are branches which
are cut/omitted, the solid path is the function to be implemented)

4.3 Exercise implementation

The object must implement active exercises, meaning that the user can move his or
her hands without help from someone else. Furthermore, this project focuses on
implementing an exercise that would be suitable for older people as they are the
biggest group of stroke survivors. The average age of stroke survivors at the time of
stroke is 70 years in men and 75 years in women [56]. However, there are also many
children who have had a stroke. Hence, although not implemented in this project, the
glove could be used to control toys. For example, the glove could act as a remote
control for steering a toy car. The interactive object can be used to perform many
hand rehabilitation exercises.
The game-playing function is great for implementing the exercises outlined in
Chapter 2.3. Different exercise movements will be mapped to different keys on the
computer. However, due to the scope of this project, and a limited number of

21
4 Proposed solution and structured design

exercises implemented, only the arrow keys, cursor movement, and mouse click will
be implemented.
The wrist flexion/extension and supination/pronation exercises, shown in Figures 2.5
and 2.6, respectively, will be used to move the cursor on the screen. Lifting the wrist
up moves the cursor upwards, left leftwards and so on. The whole arm could be used
as well to imitate the movements. Furthermore, the sensitivity should be rather low in
order to take into account imprecision and rapidness of hand movements.
The mouse click will be implemented using the finger flexion/extension and grip
strengthening exercises, shown in Figures 2.7 and 2.9, respectively. By squeezing a
ball in the hand, a mouse click will be registered. Furthermore, by measuring the
force of the grip using the sensors, the user can get instant feedback of his/her
progress.
Arrow keys will be implemented using simple finger movements as these controls
seem to be a part of many online games. Furthermore, it gives a possibility to use four
fingers simultaneously in a game and hence train them. The user has to tap a finger on
a surface, as shown for the exercise in Figure 2.8. Each finger corresponds to one
arrow key. By applying some pressure on a finger, an arrow key keystroke is
registered.
Playing computer games which are fun and make use of the keystrokes and mouse
movements mentioned above can be very beneficial for stroke rehabilitation. When
the user is motivated and enjoys playing a game, he/she is likely to perform the
exercises multiple times in a day. Although the small number of possible movements
limits the number of games the user can play. Nevertheless, there are still many
games which can be used in conjunction with the glove.

22
5 Designing, prototyping and testing

This chapter describes how the chosen idea is designed, built and tested as well as the
materials used to create the interactive object. It also gives an overview of hardware
and software implementations and of the testing phase.

5.1 Criteria and constraints

As mentioned in previous chapters, it is important to always have the end-user in
mind. Hence, the following criteria needs to be considered whilst designing.
 Weight – should be light enough to wear for prolonged periods.
 Size – should not restrict the user’s movements, and should be suitable for
children and adults.
 Easy to use – clear and easy to follow instructions.
 Motivation – is it motivating and encouraging?
 Fun factor – the user should enjoy using the device.
 Aesthetics – should fit in in an average home.
 Technical appearance – the less wires and visible components, the better.
 Adjustability – the object should take into account the user’s current abilities
and his or her progress.

5.2 Materials
A number of materials were used to create low-fidelity and high-fidelity prototypes.

5.2.1 littleBits
littleBits was used for low-fidelity prototyping. It consists of small electronic building
blocks, which can be snapped together with small magnets, such as LEDs, flex and
pressure sensors. Although their functionality is rather basic and quite restricted, it
can be used to create crude prototypes within minutes.

5.2.2 Arduino microcontrollers

The core of the interactive object is a microprocessor which controls how the device
reacts to different inputs. The microprocessors used in this project are the Arduino
UNO and the Arduino Leonardo. Arduino is great for easy and fast prototyping. It is
also inexpensive and its various boards provide with lots of functionality.
Furthermore, as both Arduino’s software and hardware are open source, there is a
large range of C libraries, wiring diagrams and extension shields available for
23
5 Designing, prototyping and testing

different types of projects, which make implementation less time-consuming and

cheaper.
The Arduino Leonardo, shown in Figure 5.1, was used for high-fidelity prototyping.
It uses its microcontroller to both run code and for USB communication with the
computer [61]. The Leonardo provides flexible communication with the computer due
to being able to handle multiple protocols. For example, the Leonardo can use USB
communications device class (USB CDC) and create a virtual serial port to the
computer. Thus, the board can act as a HID (human interface device) mouse or
keyboard, which is vital for the project. Although some other Arduino boards, such as
the UNO, were considered as well, this functionality was the deciding factor why the
Leonardo was chosen.

Figure 5.1 Arduino Leonardo [61]

The Leonardo’s operating voltage is 5V and recommended input voltage 7-12V. It
can be powered by a USB cable, but also with an AC-to-DC adapter or battery.
Hence, it could be connected to the computer wirelessly and moved around.
Additionally, since the project is heavily dependent on sensors, the number of input
and output pins is also vital. The Arduino Leonardo has 20 digital I/O pins, of which
seven can also be used as pulse-width modulated outputs and 12 as analogue input
pins. Each I/O pin can provide 40 mA of current, and hence will be sufficient to run
the sensors used in this project. The board is relatively small (68.6 mm x 53.3 mm)
and light (20 g), and hence very suitable for this project.

5.2.3 LilyPad accelerometer

The ADXL335 three axes MEMS LilyPad accelerometer, shown in Figure 5.2, is
used in this project to detect the position of the user’s hand. It has a low noise and a
low supply current of 350 µA [62], meaning it is suitable for use with the Arduino

24
 5 Designing, prototyping and testing

Leonardo. The board’s supply voltage is 5V and hence can be connected to the
Arduino’s 5V pin [63]. Furthermore, its maximum measurement range is ±3 g.

Figure 5.2 ADXL335 Accelerometer [64]

The LilyPad accelerometer is very suitable for usage in projects involving interactive
objects. It has large connection pads to make it easy to sew into clothing and can also
be washed [64]. Furthermore, it is rather small, 20 mm in diameter, and very light
[64].

5.2.4 Sensors
The flex sensors by Spectra Symbol, shown in Figure 5.3, are used in the project to
sense the bending of fingers. Its resistance when it is flat is 25𝑘Ω, and increases as it
is bent [65]. It is 2.2 inches long and hence suitable for sensing finger flexion.

25
5 Designing, prototyping and testing

Figure 5.3 Flex sensor [65]

The FSR 406 force sensing resistor by Interlink Electronics, shown in Figure 5.4, is
used to sense the force of a squeeze in low-fidelity prototypes. When no force is
applied, its resistance is very high (essentially an open circuit), but as the force
increases, the resistance decreases drastically [66].

Figure 5.4 FSR 406 force sensitive resistor [66]

However, due to its thickness and sharp edges, it was determined that it might not be
suitable for use in high-fidelity prototypes. The pressure sensitive sheet, also known
as Velostat or Linqstat, shown in Figure 5.5, was used instead. Although it needed
some preparatory work before it could be used, its thickness is only 0.2 mm and
hence much more suitable for the project [67]. The material itself consists of a
26
 5 Designing, prototyping and testing

polymeric foil and carbon black, the latter making it electrically conductive. As with
the FSR 406 described above, its resistance decreases as the force applied on it
increases.

Figure 5.5 Pressure sensitive sheet [68]

5.2.5 Other materials

Furthermore, a normal glove, where all sensors were sewn into/attached to, was used.
Two types of conductive thread were used in this project. The first one was silver-
coated thread and was used for sewing the accelerometer into the glove. Its resistance
is low over short distances and hence did not affect sensor measurements too much.
The second thread had higher resistance and was used for creating force sensitive
resistors using the pressure sensitive sheet.

5.3 Prototyping

5.3.1 Low-fidelity prototype

Low-fidelity (Lo-Fi) prototyping was used to gather quick feedback. It helps to
visualise and test the function in a fast and cheap way, and to develop ideas before
starting with high-fidelity prototyping.
littleBits was used to create the first prototype. Figure 5.6 shows a prototype which
uses two flex sensors, a force sensitive resistor and LEDs. As the user interacts with
sensors, the LEDs gradually turn on as more force is applied or the flex sensors are
bent further.

27
5 Designing, prototyping and testing

Figure 5.6 Lo-Fi prototyping with littleBits (the flex sensors are being bent by the
two blue buttons and force is applied on the force sensitive resistor)
This prototype provided with some useful information. For example, it helped to
determine which sensors should be used in the project. Furthermore, it gave an
overview of the suitability of flex and pressure sensors. It was determined that the
sensors are suitable for use in this project as they are very robust and easy to handle.
Their simplicity makes them easy to use in conjunction with an Arduino to get quick
measurements in a fast way. However, this prototype only concentrated on hardware
and not on software.
While morphing towards a high-fidelity (Hi-Fi) prototype, another Lo-Fi prototype
was created using the Arduino UNO, as shown in Figure 5.7. This step made use of
both hardware and software capabilities of the board. This prototype was used to
determine if the resistor values of the voltage dividers for the sensors are suitable, i.e.
that the sensors are not too sensitive nor insensitive. Serial() function was used to
display measurements on screen using the serial monitor. However, it was found that
it is difficult to use the UNO as an HID mouse and keyboard, and hence the Leonardo
was chosen instead, as explained in Chapter 5.2. The Flora microcontroller by
Adafruit was also considered at this stage. It is a wearable electronic platform which
can also be sewn and washed [69]. However, due to an insufficient number of input
pins, its integration into this project would have been too complex.

28
 5 Designing, prototyping and testing

Figure 5.7 Lo-Fi prototyping with the Arduino UNO

5.3.2 High-fidelity prototype

While Lo-Fi prototypes are great for fast prototyping, they lack in detail and function.
Hence, a high-fidelity (Hi-Fi) prototype was developed, which was more time-
consuming, but had complete functionality and full interactivity.
The glove can be used to simulate three different computer functionalities.
1) Arrow key keystrokes
The following algorithm performs an arrow key keystroke:
 Check if fingers are not flexed.
 Check if there is any pressure applied on any of the fingertips.
 If the conditions above are satisfied, simulate pressing the arrow key
corresponding to the fingertip.
2) Mouse click
A mouse click was simulated by using the following algorithm:
 Check if fingers are flexed.
 Check if there is any pressure applied on the palm sensor for less than
one second.
 If the conditions above are satisfied at the same time, perform a
mouse click.
3) Cursor movement

29
5 Designing, prototyping and testing

The cursor can be moved if the following conditions are satisfied:

 Check if fingers are flexed.
 Check if there is any pressure applied on the palm sensor for at least
one second.
 If the conditions above are satisfied at the same time, use the
accelerometer to measure the tilt of the hand, and move the cursor
according to the data received.
Note that it follows similar steps to the mouse click implementation.
However, a mouse click is only performed if the time of applying pressure on
the palm sensor is less than one second. However, in order to move the
cursor, the user has to keep applying force on the sensor as long as he/she
wants to move the cursor.
The top and bottom views of the final design are shown in Figures 5.8 and 5.9,
respectively.

Figure 5.8 Top view of the glove

30
 5 Designing, prototyping and testing

Figure 5.9 Bottom view of the glove

5.3.2.1 Hardware implementation

The circuit diagram of the project is shown in Appendix A.

31
5 Designing, prototyping and testing

In order to perform an arrow key keystroke, the following sensors are used: two flex
sensors on the index and middle finger, and four force sensitive resistor (FSR) on the
index, middle, ring and little fingers.
Voltage divider circuits, shown in Figure 5.10, were used for both flex and FSR
sensors. The circuit used 𝑉𝐶𝐶 = 5𝑉 input voltage, supplied by the Leonardo’s 5V pin.
Hence, the output 𝑉𝑜𝑢𝑡 is given by:
𝑅2 𝑅2
𝑉𝑜𝑢𝑡 = 𝑉𝑖𝑛 = ∗ 5𝑉. (1)
𝑅1 + 𝑅2 𝑅1 + 𝑅2

Figure 5.10 Voltage divider

In order to get a good analogue output reading from the circuit, the resistance range of
the sensors (R1 in the figure) were measured. Measurements yielded that the flex
sensors have the resistance of 𝑅1 = 25𝑘Ω when flat, and around 𝑅1 = 35𝑘Ω when a
finger is bent. As the sensors are not extremely accurate, it was important that small
changes do not influence the output much, but also that the output is sensitive to
resistance changes. Using Equation (1), it can be shown that the voltage covers the
largest range when
𝜕 𝑅2 𝑅2
( − ∗ ) = 0, (2)
𝜕𝑅2 𝑅1 + 𝑅2 𝑅1 + 𝑅2
where 𝑅1 and 𝑅1∗ are minimum and maximum resistances of the variable resistor and
𝑅2 is the value of the fixed resistor to be found. Solving Equation (2) yields that
maximum range is covered when
𝑅2 = √𝑅1 𝑅1∗ , 𝑅1 ≠ 𝑅1∗ . (3)
Thus, applying Equation (3), that the best value for 𝑅2 ≈ 29.58𝑘Ω. 𝑅2 = 27𝑘Ω was
chosen as it was the closest resistor value available. It would have been possible to
32
 5 Designing, prototyping and testing

obtain an even closer resistance using serial and parallel resistor circuits. However,
since it only makes the design more complicated, and this application is flexible on
precision, it was decided to use a single resistor.
Using Equation (1) and the resistor values above, it can be calculated that the output
voltage is approximately 2.60V when the sensor is flat and around 2.18V when a
finger is bent. Since the Leonardo use 10-bit analogue-to-digital converter, i.e.
analogue values are converted to digital values between 0 − 1023, the corresponding
output range will be between 446 − 531. Hence, 500 is a good threshold value when
determining if the sensor (finger) has been bent or not.
The force sensors have higher resistance when no pressure is applied and as more
pressure is applied, its resistance decreases. The pressure sensitive sheet was used to
handcraft the fingertip FSRs. Measurements showed that the average value of the
FSRs was about 1.5𝑘Ω when no pressure was applied, and around 600Ω when
moderate pressure was applied. Using Equation (3) yields that the output range is the
greatest when 𝑅2 ≈ 949Ω. 𝑅2 = 1𝑘Ω was chosen as it was the closest valued resistor
available. Hence, when no force is applied, the output voltage 𝑉𝑜𝑢𝑡 = 2𝑉,
corresponding to a digital value of 409, and with moderate force 𝑉𝑜𝑢𝑡 = 3.125𝑉, or
otherwise a digital value of 639. Again, the threshold value of 500 can be used to
determine whether there is any pressure on the fingertip or not.
The mouse click function uses the following sensors: two flex sensors on the index
and middle finger, and an FSR in the palm. The flex sensors are the same as those
used for the arrow key keystroke function. The FSR measures the pressure in the
palm, and acts essentially in the same way as the FSRs on the fingertips. It was
handcrafted using the pressure sensitive sheet and conductive thread. It is also worth
mentioning that a normal hand-size ball was used to apply force on the palm sensor
by squeezing the palm, as the sensor could not be conveniently reached by just
fingertips.
The resistance of the FSR in the palm with no force applied is around 23𝑘Ω, and with
force about 1.5𝑘Ω. Using Equation (3) gives that 𝑅2 ≈ 5.87𝑘Ω. The closest value
available was 𝑅2 = 5.6𝑘Ω. Using these values, 𝑉𝑜𝑢𝑡 ≈ 0.98𝑉 when there is no force
on the sensor, and 𝑉𝑜𝑢𝑡 ≈ 3.94𝑉 when force is applied. This corresponds to digital
values of 200 and 807, respectively. Hence, again, 500 is a good threshold value for
detecting force on the sensor as it lies roughly in the middle.
The cursor movement function uses the flex sensors, the FSR in the palm, and the
accelerometer. This project only uses the x and y axes of the accelerometer since the
cursor movement itself is two dimensional.
When the glove is flat on a surface, it only experiences gravitational acceleration in z
direction, i.e. 𝑥 = 𝑦 = 0 𝑔 = 0 𝑚/𝑠 2 . It was measured that the output of x and y axes
in that case was about 2.47V. When the glove is turned to the right or upwards, the
gravity of 1 𝑔 acts in x and y directions, respectively, and when turned to the left or
downwards, force of −1 𝑔 acts in x and y directions, respectively. It was measured
that the output was approximately 1.91V when 𝐹 = −1 𝑔 and about 2.98V when 𝐹 =

33
5 Designing, prototyping and testing

1 𝑔. The corresponding digital values are 505 for 2.47V, 390 for 1.91V and 610 for
2.98V.

5.3.2.2 Software implementation

The code that was used to programme the board is shown in Appendix B.
The code starts by importing the Keyboard and Mouse libraries, as shown in Figure
5.11, which are used to control the mouse and the keyboard of the computer. After
that a number of variables are defined, such as input pins and different threshold
values.

Figure 5.11 Flowchart of the program

The setup() function of the code is run only once, when the board is powered up. It
defines the pins used as input pins and also calls two functions, Mouse.begin() and
Keyboard.begin(), as shown in Figure 5.12. They start emulating the mouse and
the keyboard, respectively.

Figure 5.12 Flowchart of the setup() function

The loop() function is the central part of the code. It is a continuous function and
runs as long as the board is powered. The flowchart of the loop function is shown in
Figure 5.13.

34
 5 Designing, prototyping and testing

Figure 5.13 Flowchart of the loop() function

The function reads the output values of the flex and FSR sensors using the built-in
analogRead() function, which returns a value between 0 and 1023. These values
are then used in two other functions, mouse() and arrowkey(), which are called
inside the loop() function.
The mouse() function is shown in Figure 5.14. Its inputs are the integer values of the
flex sensors and FSR in the palm which were obtained in the loop() function. The
function starts by recording the start time of the function using the millis()
function, which returns the time in milliseconds since the program began running. It
also sets the current iteration time to zero, although technically no loop iterations
have happened at this point. It then determines if the fingers are bent and if there is
any pressure applied on the sensor in the palm. Both of these are done by comparing
the output values of the sensors with the threshold values initialised in the declaration

35
5 Designing, prototyping and testing

section of the program. If the conditions are not satisfied, the function checks if the
time difference between the start time and the time of the last iteration is less than a
threshold (1 second in the final implementation). If yes, the function uses the
Mouse.click() function which simulates a left button click and then ends.
Otherwise, the function ends instantly. However, if the sensor conditions are satisfied,
the time of the current iteration is saved. If the time difference between the start time
and the time of the current iteration is above the threshold, the mouse_distance()
function is called which returns a scaled value of the accelerometer readings. These
scaled values are then used to move the cursor on screen using the Mouse.move()
function relative to the cursor’s current location. After that, the sensor values are
updated. However, if the time is less than the period, this is done straight away. The
function then checks again if the sensors’ conditions are satisfied and continues as
described above.

Figure 5.14 Flowchart of the mouse() function

Figure 5.15 shows the flowchart of the mouse_distance() function. Its only input
is the number of the accelerometer’s pin. It starts by measuring the output of an

36
 5 Designing, prototyping and testing

accelerometer’s axis. Using the map() function, this value is then scaled down to a
number between the maximum and minimum movements desired which are defined
in the global variables’ declaration section. In this implementation, values of the x-
axis have been inverted to ensure that the cursor is moving in the right direction. If
the value is above the threshold, defined at the same place, the function returns the
value. Otherwise, the value of zero is returned. This helps to ensure that only hand
movements, which are significant enough, result in cursor movements.

Figure 5.15 Flowchart of the mouse_distance() function

The flowchart of the arrowkey() function is shown in Figure 5.16. As mentioned
above, this function is called in the loop() function and is used for simulating arrow
key keystrokes. The function starts by checking if the flex sensors’ output values are
above the threshold, i.e. the flex sensors are not bent. If it is not satisfied, the function
ends. Otherwise, the function starts to check if the output values of the FSRs are
above the threshold value, i.e. force is applied on a sensor. If yes, the corresponding
arrow key keystroke is simulated and the function continues with checking the other
sensors. Otherwise, the checks are done straight away. If all the checks have been
completed, the function ends.

37
5 Designing, prototyping and testing

Figure 5.16 Flowchart of the arrowkey() function

5.4 Testing
The interactive object was tested to get feedback about what is good and what could
be improved. This is also the basis of the iterative design process, discussed in
Chapter 3.2. The tests were also accompanied by a questionnaire that the testers filled
in, shown in Appendix C.
Firstly, a tester was presented with the pre-test questionnaire, which asked them about
their background, cognitive and motor abilities, and how often and for what purpose
they use computers. Then a test was conducted where the user tested different
functionalities of the glove. Finally, the tester filled in the post-test questionnaire
which used the system usability scale (SUS). The tester was presented with multiple

38
 5 Designing, prototyping and testing

questions which asked to assess his/her experience using the glove. He/she could also
say what they liked about the product and what could be improved.
Four people took part in the testing who were not aware of the details of the project.
All of them were in their 20s and indicated no cognitive or motor impairments in their
responses. Three of the users were right-handed and one left-handed. Although the
glove was designed for right-handed people, he was willing to use his right hand and
was successfully able to interact with the glove. All of them are also frequently using
computers for different activities and hence had no difficulty understanding the
functions of the glove.
Firstly, the testers were given a short summary about the project and the glove. After
learning about different functions of the glove and filling in the pre-test questionnaire,
they could go online and pick a game which they liked and could be played using the
functions of the glove. For example, one tester successfully played a game called
Flappy Bird which used only mouse clicks as its input to guide a bird through rows of
pipes. Furthermore, the testers also tried to control the computer using the glove, e.g.
opening different programs and surfing the web.
In order to follow the iterative design process, after the first two tests, the glove was
adapted to the feedback received from the testers. A few recommendations were
suggested. Firstly, in the initial Hi-Fi design, the sensors were just attached to the
glove by securing them in place with thread. However, during testing both testers
experienced that sensors got stuck in it when they were moved, especially the flex
sensors. As a solution they were covered by a fabric which was then sewn into the
glove, as can be seen from Figures 5.8 and 5.9. The fabric is stretchable and hence
very suitable for this application, where the shape of it needs to change as a result of
bending. Secondly, there were some complications using the glove, as its response
was either too sensitive or insensitive. The resistor values were recalculated, replaced
and the threshold values in the code changed according to the values mentioned in
Chapter 5.3.2.1. Lastly, both testers also recommended to sew the cables into the
glove as they were too loose at this point of testing.
After these tests, the design was modified to take into account the recommendations.
After that, two new testers tested the glove and filled in the post-test questionnaire
with their feedback. They reported a problem with using the glove to move the cursor,
something which had been already mentioned during the first two tests. Although it
was better this time, the cursor movement could not be used to play games due to its
slow and a bit imprecise movements. Furthermore, during one testing the tester
revealed that the ball used in the testing was too stiff as the hand became tired
quickly. A softer ball was used instead which was easier to use for longer periods.
However, it should be noted that extremely soft balls are not recommended as they
might not apply enough force on the palm sensor. Hence, a mouse click might not be
registered, unless changing the threshold values. Lastly, during all tests, people were
afraid of damaging the sensors, especially the flex sensors, when putting on or taking
off the glove. Hence, it was recommended to do it slowly in a way that the sensors are
not bent too much. Unfortunately, it was not possible to change this due to inevitable
characteristics of the sensors.
39
5 Designing, prototyping and testing

Although due to the low number of testers no solid conclusions can be made, all
testers said that they enjoyed using the glove and would also do so in the future. They
also said that the glove was easy to use and its low number of functions did not
require much learning. Apart from having a few problems with using the glove to
move the cursor, they said that using the glove does not require much help from
others. The feedback also revealed that although users with different hand sizes could
use the glove itself, its material might not be the best for longer periods of use as the
material is rather thick. Furthermore, although the testers did not think that the glove
itself was aesthetically unpleasing, they said that it would look nicer if the cables
were less visible. Additionally, they also thought that having more functions would be
beneficial.
All the users said that they believe the glove could be used to motivate stroke
rehabilitees. They suggested that in addition to playing games, it could be used to
perform other activities such as browsing the Internet. However, their main
motivation was that they enjoyed using the glove and when playing games, they
performed different hand movements for multiple times by concentrating on the game
rather than on the hand movements itself. Hence, it would be likely the case with
many other people. Furthermore, it was mentioned that using the glove to play games
is also competitive and hence motivating. For example, the testers wanted to beat
their own and other testers’ high scores by playing games over and over again. Hence,
this might also be encouraging for competitive stroke rehabilitees.
In conclusion, the testers thought that the glove was easy to use and could be
beneficial for exercising hand muscles, for example those involved in gripping.
Although there are things which could be improved, it was generally functioning
well.

40
6 Conclusions

This chapter includes the main findings of the project. It also analyses limitations of
the final design which could be improved upon and also provides recommendations
for future work.

6.1 Discussion and conclusions

Stroke as a disease affects millions of people around the world. Although it is mainly
prevalent in developed countries, the number of stroke survivors is also increasing in
developing countries. Those who survive a stroke have generally lost some of their
ability to perform activities of daily life. Hence, it is highly recommended that in
addition to rehabilitation with a physician the stroke survivor would exercise at home
as well. However, many studies have shown that repeating the same exercises every
day without seeing any improvement may be very demotivating as well as
discouraging. Thus, many people give up rehabilitation completely due to their lack
of motivation.
A solution is to create interactive objects which would take into account a person’s
current abilities as well as interests. Several studies have shown that those using
interactive objects tend to be more motivated and show better results than those
relying on conventional methods. However, the options are rather limited and some
products are directed more on healthcare institutions rather than on individual people.
Thus, it is important to develop interactive objects which would not be expensive, but
would be motivating and would cater for a wide range of interests. It is also important
to take into account other factors such as how easy is the device to use and how well
does it fit in with the aesthetics at home. Additionally, the user should get some
feedback, for example regarding his/her progress.
This thesis was concerned with how interactive objects can be beneficial for mobility
training and stroke rehabilitation. A number of ideas were proposed and their
advantages and disadvantages analysed. It was decided that using a glove, which
could be used to control a computer’s mouse and keyboard, could be a good approach
as it can be used to play online games and is suitable for implementation in the short
timeframe of the project.
After studying previous papers, it was determined that the final product should be
easy to use, lightweight, small in size, aesthetically pleasing, and most importantly
fun and motivating. Furthermore, the glove also caters for a wide range of stroke
survivors, not for just a particular age group.

41
6 Conclusions

The glove is used to perform different hand and finger exercises. For example, in
order to move the cursor, the wrist pronation and supination exercise must be
performed. Additionally, a grip strengthening exercise is used to perform a mouse
click while finger lifting is used to simulate arrow key keystrokes.
The Arduino Leonardo microcontroller was used as it can be used as a human
interface device to emulate a mouse and keyboard. Furthermore, it is small in size,
light and can be used for many different purposes. Both Arduino’s software and
hardware are open sources and hence the resources are vast. In addition to the board,
an accelerometer, flex sensors and force sensitive resistors were used to determine
hand and finger movements. Thus, the production costs of the glove are much cheaper
compared to other alternatives on the market. This in turn means that the product
could be directed at individuals rather than at healthcare institutions.
An iterative design process was used to keep the end user in mind. This means that
after designing and prototyping, tests were done and the results were used to change
the design so that it would be more comfortable to use. The system usability scale
was used to get feedback about the object in a simple, fast and inexpensive way.
Although stroke survivors were not involved in the testing phase, it nevertheless gave
useful feedback which could be used to improve general functionality. Four testers
tested the glove and all of them reported that the glove was easy and enjoyable to use
and could possibly be motivating for those who have suffered from a stroke and
might need help with recovering abilities of their hand. They claimed that the main
advantage of using the glove is that the concentration is on playing games and
interacting with the glove rather than performing exercises.
As the technology-assisted rehabilitation is becoming more popular, it is important to
have discussions with the target group to determine their expectations as well as
concerns. Furthermore, it is also important that stroke survivors would be confident in
using the device. This could be solved by splitting complex tasks down into simpler
ones, for example, which is used by the technology-supported task-orientated arm
training method. Additionally, it is also important to create tasks which have a
meaning, i.e. the rehabilitee needs to do the same task in his/her daily life. However,
there should be a balance between rehabilitation and what the user likes to do to keep
him/her motivated. Furthermore, too much exercising might be discouraging as well.
Thus, involving stroke patients is vital in order to create devices which would be both
motivational as well as beneficial.
In conclusion, interactive objects can be very supportive when it comes to stroke
rehabilitation. In addition to helping the users to perform exercises in a correct way, it
can also be motivating and encouraging. The number of solutions on the market is
constantly growing and the research focus on the topic is becoming more important. It
is likely that in the future rehabilitation would involve interactive objects which
would take into account the user’s interests and abilities and hence would help with
performing activities of daily life.

42
 6 Conclusions

6.2 Future work

Due to the scope of the project, limitations of the final product are inevitable. Hence,
the following could be improved to make the glove even more usable.
Firstly, the current prototype lacked adaptability by not having an option to modify
threshold values of the functions without re-programming the microprocessor. Hence,
a functionality, which could determine the user’s abilities such as grip strength, would
be advantageous so the same product could be used by different people without any
programming intervention. Furthermore, it would also be good if people could track
their own performance, for example by collecting data which could later be used to
create graphs.
Since some sensors of the glove were handcrafted, they were not as precise as those
manufactured by a specialised company. Hence, the sensors in the current design
could be swapped. Although precision is not extremely important in this application,
it might ensure better overall functionality of the glove. Furthermore, the cursor
movement function and sensitiveness needs to be improved to make it more accurate
and easier to use.
Although the Arduino Leonardo board is great for prototyping, its size and
functionalities are not the most suitable for use in the final product. Instead, a
microcontroller chip should be used which might result in more complex software
and hardware implementations, but greatly reduces the size of the final product.
Another improvement would be to emulate more mouse and keyboard functions.
Currently, only the mouse click, cursor movement and arrow key functions are
available. However, with addition of more sensors, it would also be possible to use
other keys on the keyboard. Furthermore, it would be an advantage if the glove could
be used to control other devices, such as the TV, as well.
It is also important to create online games which would be directed at stroke survivors
and would allow them to use their hands for various exercises. A large database of
games would potentially have something for everybody, for example games for
different age groups or people with different interests.
The aesthetics of the glove could also be improved as the current design did not make
much effort to hide the cables. Furthermore, creating a wireless version of the glove
would also decrease the number of cables as well as would provide extra freedom.
Additionally, the glove could be made of a material which might be more user-
friendly when using for extended periods of time, for example not as thick.
Furthermore, the sensors should be secured in a way that they could not be damaged
easily, for example when putting on or taking off the glove, which the testers were
worried about when testing the high-fidelity prototypes.
The final prototype was made for the use of the right hand. However, as the weaker
hand of stroke survivors is dependent on the hemisphere of the brain where the stroke
occurred, there are also many people who need rehabilitation with their left hand.
Hence, a similar product, but for the left hand, should also be created.

43
6 Conclusions

Finally, the testing did not involve any stroke survivors as in previous ActivABLES
projects. Hence, in the future it would be very advantageous to involve them in every
stage of the project, so the final product would be designed according to actual needs.

44
References

[1] V. L. Feigin, R. V. Krishnamurthi, P. Parmar, B. Norrving, G. A. Mensah, D. A.

Bennett, S. Barker-Collo, A. E. Moran, R. L. Sacco, T. Truelsen, S. Davis, J. D.
Pandian, M. Naghavi, M. H. Forouzanfar, G. Nguyen, C. O. Johnson, T. Vos, A.
Meretoja, C. J. Murray and G. A. Roth, “Update on the Global Burden of
Ischemic and Hemorrhagic Stroke in 1990-2013: The GBD 2013 Study,”
Neuroepidemiology, vol. 45, no. 3, pp. 161-176, 2015.
[2] American Stroke Association, “Understanding Stroke Risk,” [Online]. Available:
http://www.strokeassociation.org/STROKEORG/AboutStroke/UnderstandingRi
sk/Understanding-Stroke-Risk_UCM_308539_SubHomePage.jsp. [Accessed 10
April 2016].
[3] World Heart Federation, “Stroke Incidence and Mortality Rates Found to be
Higher in Developing than in Developed Countries for First Time,” 19 June
2010. [Online]. Available: http://www.world-heart-federation.org/what-we-
do/congresses/world-congress-of-cardiology-scientific-sessions-
2010/press/press-releases/detail/article/stroke-incidence-and-mortality-rates-
found-to-be-higher-in-developing-than-in-developed-countries-fo. [Accessed 10
April 2016].
[4] N. Poungvarin, “Stroke in the developing world,” The Lancet, vol. 352, pp. 19-
22, 1998.
[5] Stroke Association, “What is a stroke?,” February 2015. [Online]. Available:
https://www.stroke.org.uk/what-stroke/what-stroke. [Accessed 24 March 2016].
[6] Stroke Association, “Ischaemic stroke,” February 2015. [Online]. Available:
https://www.stroke.org.uk/what-stroke/what-stroke/ischaemic-stroke. [Accessed
24 March 2016].
[7] American Heart Association, “Ischemic Strokes (Clots),” 28 March 2016.
[Online]. Available: http://www.strokeassociation.org/STROKEORG/AboutStro
ke/TypesofStroke/IschemicClots/Ischemic-Strokes-
Clots_UCM_310939_Article.jsp. [Accessed 29 March 2016].
[8] Stroke Association, “Haemorrhagic stroke,” February 2015. [Online]. Available:
https://www.stroke.org.uk/what-stroke/what-stroke/haemorrhagic-stroke.
45
References

[Accessed 24 March 2016].

[9] National Heart, Lung, and Blood Institute (NHLBI), “What Is a Stroke?,” 28
October 2015. [Online]. Available: http://www.nhlbi.nih.gov/health/health-
topics/topics/stroke. [Accessed 10 March 2016].
[10] Central and North West London NHS Foundation Trust, “Caring for your heart,”
February 2014. [Online]. Available: http://www.cnwl.nhs.uk/wp-
content/uploads/CNWL_OHS_Caring_for_your_heart_booklet.pdf. [Accessed
10 March 2016].
[11] Allina Health, “What is a stroke?,” 12 September 2011. [Online]. Available:
http://www.allinahealth.org/Health-Conditions-and-Treatments/Health-
library/Patient-education/Understanding-Stroke/Stroke-overview/What-is-a-
stroke/. [Accessed 10 April 2016].
[12] Mayo Clinic, “Stroke rehabilitation: What to expect as you recover,” 11 June
2014. [Online]. Available: http://www.mayoclinic.org/stroke-rehabilitation/art-
20045172. [Accessed 11 March 2016].
[13] Allina Health, “The goal of stroke rehabilitation,” 12 September 2011. [Online].
Available: http://www.allinahealth.org/Health-Conditions-and-
Treatments/Health-library/Patient-education/Understanding-Stroke/Recovery-
and-rehabilitation/The-goal-of-stroke-rehabilitation/. [Accessed 4 April 2016].
[14] National Institute of Neurological Disorders and Stroke, “Post-Stroke
Rehabilitation,” National Institutes of Health, September 2014. [Online].
Available: http://stroke.nih.gov/materials/rehabilitation.htm. [Accessed 20
March 2016].
[15] G. Alankus, R. Proffitt, C. Kelleher and J. Engsberg, “Stroke Therapy through
Motion-Based Games: A Case Study,” ACM Transactions on Accessible
Computing (TACCESS), vol. 4, no. 1, 2011.
[16] M. F. Levin, J. H. Burridge, S. K. Subramanian and J. M. Solomon, “Stroke: new
technologies (FS-24),” World Confederation for Physical Therapy, 25 June
2015. [Online]. Available: http://www.wcpt.org/congress/fs/86. [Accessed 11
March 2016].
[17] National Stroke Association, “HOPE: The Stroke Recovery Guide,” 2010.
[Online]. Available: http://www.stroke.org/sites/default/files/resources/NSA-
Hope-Guide.pdf. [Accessed 10 April 2016].
[18] Medical Advisory Secretariat, Health Quality Ontario, “Constraint-Induced
Movement Therapy for Rehabilitation of Arm Dysfunction After Stroke in
Adults: An Evidence-Based Analysis.,” Ontario Health Technology Assessment
Series, vol. 11, no. 6, pp. 1-58, 2011.
[19] D. T. Wade and R. L. Hewer, “Functional abilities after stroke: measurement,
46
 References

natural,” Journal of Neurology, Neurosurgery, and Psychiatry, vol. 50, no. 2, pp.
177-182, 1987.
[20] F. I. Mahoney and D. W. Barthel, “Functional evaluation: The Barthel Index,”
Maryland State Medical Journal, vol. 14, pp. 61-65, 1965.
[21] M. Shaughnessy and B. M. Resnick, “Testing a Model of Post-Stroke Exercise
Behavior,” Rehabilitation Nursing, vol. 31, no. 1, pp. 15-21, 2006.
[22] L. Axelrod, G. Fitzpatrick, J. Burridge, S. Mawson, P. P. Smith, T. Rodden and
I. Ricketts, “The reality of homes fit for heroes: design challenges for
rehabilitation technology at home,” Journal of Assistive Technologies, vol. 3, no.
2, pp. 35-43, 2009.
[23] C. Bjørkquist, H. Ramsdal and R. Kjetil, “User participation and stakeholder
involvement in health care innovation – does it matter?,” European Journal of
Innovation Management, vol. 18, no. 1, pp. 2-18, 2015.
[24] G. Fitzpatrick, M. Balaam and S. Rennick Egglestone, “Involving Stroke
Survivors in Designing for Rehabilitation,” in Therapeutic Strategies: A
Challenge for User Involvement in Design, Reykjavik, 2010.
[25] J. Curtis, L. Ruijs, M. de Vries, R. Winters and J.-B. Martens, “Rehabilitation of
handwriting skills in stroke patients using interactive games: a pilot study,” in
CHI '09 Extended Abstracts on Human Factors in Computing Systems, Boston,
2009.
[26] M. Balaam, S. Rennick Egglestone, G. Fitzpatrick, T. Rodden, A.-M. Hughes, A.
Wilkinson, T. Nind, L. Axelrod, E. Harris, I. Ricketts, S. Mawson and J.
Burridge, “Motivating Mobility: Designing for Lived Motivation in Stroke
Rehabilitation,” in Proceedings of the SIGCHI Conference on Human Factors in
Computing Systems, Vancouver, 2011.
[27] F. Delbressine, A. Timmermans, L. Beursgens, M. de Jong, A. van Dam, D.
Verweij, M. Janssen and P. Markopoulos, “Motivating arm-hand use for stroke
patients by serious games,” in 34th Annual International Conference of the IEEE
EMBS, San Diego, 2012.
[28] S. M. Michaelsen, S. Jacobs, A. Roby-Brami and M. F. Levin, “Compensation
for distal impairments of grasping in adults with hemiparesis,” Experimental
Brain Research, vol. 157, no. 2, pp. 162-173, 2004.
[29] A. A. A. Timmermans, R. P. J. Geers, J. A. Franck, P. Dobbelsteijn, A. I. F.
Spooren, H. Kingma and H. A. M. Seelen, “T-TOAT: A method of task-oriented
arm training for stroke patients suitable for implementation of exercises in
rehabilitation technology,” in IEEE International Conference on Rehabilitation
Robotics, Kyoto, 2009.
[30] A. A. A. Timmermans, H. A. M. Seelen, R. P. J. Geers, P. K. Saini, S. Winter, J.
47
References

te Vrugt and H. Kingma, “Sensor-Based Arm Skill Training in Chronic Stroke

Patients: Results on Treatment Outcome, Patient Motivation, and System
Usability,” IEEE Transactions on Neural Systems and Rehabilitation
Engineering, vol. 18, no. 3, pp. 284-292, 2010.
[31] Certec, “Activables,” Lunds universitet, 2 December 2015. [Online]. Available:
http://www.certec.lth.se/forskning/activables/. [Accessed 19 April 2016].
[32] ActivABLES, “About,” [Online]. Available: http://www.activables.org/about/.
[Accessed 19 April 2016].
[33] C. Magnusson, H. Caltenco, M. Kyto and et al., “Integrating Effective Stroke
Rehabilitation into Everyday Life through Tangible Interaction,” in Proceedings
of the 9th Nordic Conference on Human-Computer Interaction: Game Changing
Design (NordiCHI’16), 2016.
[34] National Institute of Neurological Disorders and Stroke, “Post-Stroke
Rehabilitation,” National Institutes of Health , September 2014. [Online].
Available: http://stroke.nih.gov/materials/rehabilitation.htm. [Accessed 20
March 2016].
[35] Neofect, “RAPAEL Smart Glove™,” [Online]. Available:
http://www.neofect.com/en/wp-content/uploads/2016/03/NEOFECT-USA-
RAPAEL-Smart-Glove.pdf. [Accessed 21 April 2016].
[36] J.-H. Shin, M.-Y. Kim, J.-Y. Lee, Y.-J. Jeon, S. Kim, S. Lee, B. Seo and Y.
Choi, “Effects of virtual reality-based rehabilitation on distal upper extremity
function and health-related quality of life: a single-blinded, randomized
controlled trial,” Journal of NeuroEngineering and Rehabilitation, vol. 13, no.
17, 2016.
[37] Neofect, “RAPAEL Smart Glove for Hand Rehab,” [Online]. Available:
http://www.neofect.com/en/product/rapael/. [Accessed 21 April 2016].
[38] Forbes, “The Rapael Smart Glove,” [Online]. Available: http://blogs-
images.forbes.com/jenniferhicks/files/2015/11/RAPAEL-Smart-
Glove_NEOFECT-1-1200x799.png. [Accessed 21 April 2016].
[39] Bioness Inc., “Products,” [Online]. Available:
http://www.bioness.com/Products.php. [Accessed 23 April 2016].
[40] Bioness Inc., “About Us,” [Online]. Available:
http://www.bioness.com/About_Us.php. [Accessed 23 April 2016].
[41] R. C. Senelick, “Technological Advances in Stroke Rehabilitation— High Tech
Marries High Touch,” US Neurology, vol. 6, no. 2, pp. 102-104, 2010.
[42] Bioness Inc., “How Does It Work?,” [Online]. Available:
http://www.bioness.com/Products/L300_for_Foot_Drop/How_Does_It_Work.

48
 References

php. [Accessed 27 April 2016].

[43] Bioness Inc., “What is it?,” [Online]. Available:
http://www.bioness.com/Products/L300_Plus_For_Thigh_Weakness/What_is_
It.php. [Accessed 27 April 2016].
[44] Bioness Inc., “H200 for Hand Paralysis,” [Online]. Available:
http://www.bioness.com/Healthcare_Professionals/Exoskeletal_Products/
H200_for_Hand_Paralysis.php. [Accessed 27 April 2016].
[45] J. M. Hausdorff and H. Ring, “The effect of a new lower-limb neuroprosthesis
on physical and social functioning,” Journal of Neurologic Physical Therapy,
vol. 30, no. 4, p. 207, 2006.
[46] . J. M. Hausdorff and H. Ring, “Effects of a new radio frequency-controlled
neuroprosthesis on gait symmetry and rhythmicity in patients with chronic
hemiparesis,” American Journal of Physical Medicine & Rehabilitation, vol. 87,
no. 1, pp. 4-13, 2008.
[47] Unicare Ltd., “Bioness L300,” [Online]. Available:
http://www.unicare.com.mt/Images/bioness%20l300.png. [Accessed 2016 April
27].
[48] 360 O&P, [Online]. Available:
http://www.360oandp.com/Data/Sites/1/productImages/325/Product-Bioness-
Ness-H200P3.jpg. [Accessed 27 April 2016].
[49] G. Saposnik, R. Teasell, M. Muhammad, J. Hall, W. McIlroy, D. Cheung, K. E.
Thorpe, L. G. Cohen and M. Bayley, “Effectiveness of Virtual Reality Using Wii
Gaming Technology in Stroke Rehabilitation: A Pilot Randomized Clinical Trial
and Proof of Principle,” Stroke, vol. 41, no. 7, pp. 1477-1484, 2010.
[50] J. Jensen, “Reclaim Your Strength With Arm Exercises For Stroke Recovery,”
12 October 2015. [Online]. Available: http://www.saebo.com/reclaim-your-
strength-with-arm-exercises-for-stroke-recovery/. [Accessed 29 April 2016].
[51] Stroke-Rehab.com, “Passive Range of Motion,” [Online]. Available:
http://www.stroke-rehab.com/passive-range-of-motion.html. [Accessed 29 April
2016].
[52] Pinterest, “Hand and Finger Exercises,” [Online]. Available:
https://www.pinterest.com/pin/300333868870750820/. [Accessed 10 May 2016].
[53] Geared for fitness, “3 Silicone Hand Strengthener Exercise Rings – Grip
Strengtheners for Computer Users, Physical Hand Therapy & Resistance
Training plus Rehabilitation of Muscles for Stroke, RSI & Stress Relief,”
[Online]. Available: https://www.gearedforfitness.com/product/3-silicone-hand-
strengthener-exercise-rings-grip-strengtheners-for-computer-users-physical-
hand-therapy-resistance-training-plus-rehabilitation-of-muscles-for-stroke-rsi-
49
References

stress-relief/. [Accessed 29 April 2016].

[54] E. Hornecker and J. Buur, “Getting a grip on tangible interaction: a framework
on physical space and social interaction,” in Proceedings of the SIGCHI
Conference on Human Factors in Computing Systems, New York, 2006.
[55] J. Nielsen, “Guerrilla HCI: Using Discount Usability Engineering to Penetrate
the Intimidation Barrier,” Nielsen Norman Group, 1 January 1994. [Online].
Available: https://www.nngroup.com/articles/guerrilla-hci/. [Accessed 17 April
2016].
[56] A. Cooper, R. Reimann and D. Cronin, About Face 3: The Essentials of
Interaction Design, Indianapolis: Wiley Publishing, Inc., 2007.
[57] Y. Rogers, H. Sharp and J. Preece, Interaction design: beyond human-computer
interaction, Chichester: John Wiley & Sons Ltd, 2011.
[58] C.-M. Karat, “Cost-benefit analysis of iterative usability testing,” INTERACT '90
Proceedings of the IFIP TC13 Third Interational Conference on Human-
Computer Interaction, pp. 351-356, 1990.
[59] U.S. Department of Health & Human Services, “System Usability Scale (SUS),”
[Online]. Available: http://www.usability.gov/how-to-and-tools/methods/system-
usability-scale.html. [Accessed 15 April 2016].
[60] Stroke Association, “Can a movement-sensing wristwatch prompt arm
rehabilitation exercise at home?,” 5 January 2015. [Online]. Available:
https://www.stroke.org.uk/research/can-movement-sensing-wristwatch-prompt-
arm-rehabilitation-exercise-home. [Accessed 29 April 2016].
[61] Arduino, “Arduino Leonardo,” [Online]. Available:
https://www.arduino.cc/en/Main/ArduinoBoardLeonardo. [Accessed 15 May
2016].
[62] Analog Devices, ADXL335 datasheet, 2009.
[63] SparkFun, “LilyPad Accelerometer Schematic,” 22 November 2010. [Online].
Available: http://cdn.sparkfun.com/datasheets/Dev/LilyPad/LilyPad-
Accelerometer-v20.pdf. [Accessed 20 May 2016].
[64] SparkFun Electronics, “LilyPad Accelerometer ADXL335,” [Online]. Available:
https://www.sparkfun.com/products/9267. [Accessed 24 May 2016].
[65] electroSome, “Flex Sensor 2.2",” [Online]. Available:
https://electrosome.com/shop/flex-sensor-2-2/. [Accessed 15 May 2016].
[66] Interlink Electronics, FSR 406 datasheet.
[67] Electrokit Sweden AB , “Tryckkänslig ledande plast (Velostat/Linqstat),”
[Online]. Available: http://www.electrokit.com/tryckkanslig-ledande-plast-
50
 References

velostat-linqstat-280x280mm.51818. [Accessed 16 May 2016].

[68] C. Perabo, “LINQSTAT Volume Conductive Film for EMF Shielding,” 3 March
2008. [Online]. Available: http://www.caplinq.com/blog/linqstat-volume-
conductive-film-for-emf-shielding_37/. [Accessed 16 May 2016].
[69] Adafruit, “FLORA - Wearable electronic platform: Arduino-compatible,”
[Online]. Available: https://www.adafruit.com/product/659. [Accessed 17 May
2016].

51
Appendix A: Schematic

53
Appendix B: Code

/*
* Title: Glove
* Author: Mihkel Hiob
* Date: Jun 2016
* Code version: 2.3
*/

#include <Keyboard.h>
#include <Mouse.h>

// Pin definitions
const int flex_index = A0; // Index finger, flex sensor
const int flex_middle = A1; // Middle finger, flex sensor
const int fsr_palm = A2; // Palm, pressure sensor
const int acc_x = A3; // Accelerometer, x-axis
const int acc_y = A4; // Accelerometer, y-axis
const int fsr_index = A8; // Index finger, pressure sensor
const int fsr_middle = A9; // Middle finger, pressure sensor
const int fsr_ring = A10; // Ring finger, pressure sensor
const int fsr_pinky = A11; // Pinky finger, pressure sensor

// Threshold values
const int click_thres = 500; // Mouse click threshold
const int flex_thres = 500; // Flex sensors threshold
const int keystroke_thres = 500; // Keystroke threshold
(finger FSRs)

const long period = 1000; // Duration before the palm

pressure sensor must be activated in order to move the cursor
(in ms)

// Accelerometer and cursor movement values

const int max_move = 20; // Max movement of cursor (in
pixels)
const int move_thres = max_move/5; // Minimum movement (in
pixels)
const int minimum = 390; // Min value of the accelerometer
reading
const int maximum = 610; // Max value of the accelerometer
reading
55
Appendix B: Code

void setup() {
pinMode(flex_index, INPUT);
pinMode(flex_middle, INPUT);
pinMode(fsr_palm, INPUT);
pinMode(acc_x, INPUT);
pinMode(acc_y, INPUT);
pinMode(fsr_index, INPUT);
pinMode(fsr_middle, INPUT);
pinMode(fsr_ring, INPUT);
pinMode(fsr_pinky, INPUT);

Mouse.begin(); // Initialise mouse control

Keyboard.begin(); // Initialise keyboard control

void loop() {
int fsr_palm_val; // Palm sensor’s value
int flex_index_val, flex_middle_val; // Flex sensors' values
int fsr_index_val, fsr_middle_val, fsr_ring_val,
fsr_pinky_val; // Finger FSR sensors' values

// Sensor measurements
fsr_palm_val = analogRead(fsr_palm);
flex_index_val = analogRead(flex_index);
flex_middle_val = analogRead(flex_middle);
fsr_index_val = analogRead(fsr_index);
fsr_middle_val = analogRead(fsr_middle);
fsr_ring_val = analogRead(fsr_ring);
fsr_pinky_val= analogRead(fsr_pinky);

mouse(fsr_palm_val, flex_index_val, flex_middle_val); //

Mouse functions
arrowkey(fsr_index_val, fsr_middle_val, fsr_ring_val,
fsr_pinky_val, flex_index_val, flex_middle_val); // Arrow key
keystroke function

/*
* Simulates arrow key keystrokes if the flex sensors are not
* bent and the pressure on a finger FSR is at least the
* threshold value.
*
* fsr1_val, fsr2_val, fsr3_val, fsr4_val: measured values of
the finger pressure sensors
* flex1_val, flex2_val: measured values of the flex sensors
*/
void arrowkey(int fsr1_val, int fsr2_val, int fsr3_val, int
fsr4_val, int flex1_val, int flex2_val) {
if ((flex1_val > flex_thres) && (flex2_val > flex_thres)) {
56
 Appendix B: Code

if (fsr1_val >= keystroke_thres) {

Keyboard.write(KEY_LEFT_ARROW);
}

if (fsr2_val >= keystroke_thres) {

Keyboard.write(KEY_UP_ARROW);
}

if (fsr3_val >= keystroke_thres) {

Keyboard.write(KEY_RIGHT_ARROW);
}

if (fsr4_val >= keystroke_thres) {

Keyboard.write(KEY_DOWN_ARROW);
}

delay(100);
}
}

/*
* Performs a mouse click if the flex sensors are bent and the
* pressure on the palm has been at least the threshold value
* for less than the period. If at least the period, moves the
* cursor.
*
* fsr_val: measured value of the palm sensor
* flex1_val, flex2_val: measured values of the flex sensors
*/
void mouse(int fsr_val, int flex1_val, int flex2_val) {
unsigned long starttime = millis(); // Register the starting
time of the function
long previoustime = 0; // Last registered time
int xDist, yDist;

while ((flex1_val <= flex_thres) && (flex2_val <= flex_thres)

&& (fsr_val >= click_thres)) {
previoustime = millis();

// Enable mouse movement

if (previoustime - starttime >= period) {
xDist = mouse_distance(acc_x);
yDist = mouse_distance(acc_y);
Mouse.move(xDist, yDist, 0); // Move cursor
delay(10);
}

// Update the values

fsr_val = analogRead(fsr_palm);
flex1_val = analogRead(flex_index);
flex2_val = analogRead(flex_middle);
}
57
Appendix B: Code

// The pressure was applied for less than 1 second (period)

if ((previoustime - starttime < period) && (previoustime >
0)) {
Mouse.click(MOUSE_LEFT);
}
}

/*
* Reads the value from the accelerometer and determines
* how much to move the cursor and returns this distance.
*
* pin_number: measurement pin
*/
int mouse_distance(int pin_number) {
int distance = 0;
int value = analogRead(pin_number);
int scaled_value = map(value, minimum, maximum, -max_move,
max_move);

// Check if the scaled value is above the threshold

if (abs(scaled_value) >= move_thres) {
distance = scaled_value;
}

// Invert x-axis values to move the cursor in correct

direction
if (pin_number == acc_x) {
distance = -distance;
}

return distance;
}

58
Appendix C: Questionnaire

Pre-test questionnaire
A. Basic information
1. Gender:  Male  Female
2. Age ____ years
B. Cognitive and motor abilities
1. How would you rate your cognitive and motor abilities?
 Very good
 Good
 Moderate
 Poor
 Severely affected
2. Do you have any kind of cognitive or motor impairments? If yes, please
provide with some details.  Yes  No
______________________________________________________________
______________________________________________________________
______________________________________________________________
3. Dominant hand:  Left  Right
C. Experience with computers
1. Do you own a computer?  Yes  No
2. If yes, how long have you used a computer?
 Less than a year
 1-3 years
 More than 3 years
3. On average, how many times do you use a computer in a week for the
following activities?
0 1-5 6-10 More
than 10
Browsing the Internet
Writing
Playing games
Other (please specify)

59
Appendix C: Questionnaire

Post-test questionnaire (System usability scale adapted from [59]):

A. System usability scale
1 – strongly disagree 5 – strongly agree
1 2 3 4 5
1. I think that I would like to use this glove frequently.     
2. I found the glove unnecessarily complex.     
3. I thought the glove was easy to use.     
4. I think that I would need the support of a technical     
person to be able to use this glove.
5. I found the various functions in this glove were well     
integrated.
6. I thought there was too much inconsistency in this     
glove.
7. I would imagine that most people would learn to use     
this glove very quickly.
8. I found the glove very cumbersome to use.     
9. I felt very confident using the glove.     
10. I needed to learn a lot of things before I could get     
going with this glove.

B. Feedback
What did you like about using the glove?
_____________________________________________________________________
_____________________________________________________________________
_____________________________________________________________________
What do you think could be improved?
_____________________________________________________________________
_____________________________________________________________________
_____________________________________________________________________
Do you think that the glove is fun to use and could potentially motivate those who
have suffered from a stroke? Please explain your answer.  Yes  No
_____________________________________________________________________
_____________________________________________________________________
_____________________________________________________________________

60