SensorShoe: Mobile Gait Analysis

SensorShoe: Mobile Gait

Analysis for Parkinson's
Disease Patients
- User Interface Design Report -

P.J.M. Havinga
R.G. Kauw-A-Tjoe
M. Marin-Perianu
J.P. Thalen

Faculty of Engineering Technology

1
SensorShoe: Mobile Gait Analysis

Faculty of Electrical Engineering, Mathematics and Computer Science

University of Twente

2
SensorShoe: Mobile Gait Analysis

Table of Contents
^Introduction ..................................... 3 4.2.Shoe Module ........................................ 21
Sensors ................................................................... 21
1.1. Motivation............................................ 3 Communication ........................................................ 21
1.2.Objectives .............................................. 3 4.3. PDA Module ...................................... 22
1.3. Focus of this Study .............................. 4 Specifications ............................................................ 22
Assumptions / Limitation .............................................. 4 Functionality. ............................................................. 22
1.4. Approach............................................. 4 Concept Evaluation & Selection ................................ 27
Result .......................................................................... 4 Detailed Interface Design .......................................... 28
4.4. Feedback Module ............................. 29
2. Literature Research ..................... 5 Feedback Requirements ........................................... 29
2.1. Parkinson's Disease ............................ 5 Mechanical Actuators ................................................ 30
Causes ........................................................................ 5 Actuator Implementation ........................................... 32
Symptoms ................................................................... 5 4.5. Conclusions ...................................... 35
Treatments .................................................................. 6
2.2. Gait Analysis ....................................... 7 5. Evaluation & Recommendations
2.3. Gait Feedback Cues ........................... 8 36
Visual Cues ................................................................. 8 5.1. Evaluation ......................................... 36
Audio Cues .................................................................. 9 System Concept Development .................................. 38
Somatosensory Cues .................................................. 9 Interface Development .............................................. 38
2.4. Conclusions ...................................... 10 5.2. Usability Test .................................... 39
GUI Prototype ........................................................... 39
3. Concept Development ............... 11 Test Setup................................................................. 40
3.1. User Characteristics .......................... 12 5.3. Conclusions ...................................... 42
PD Patient ................................................................. 12
Caregiver. .................................................................. 12 6. Literature ................................... 43
3.2. Use Scenario .................................... 13 6.1............................................................. Pa
Persona's .................................................................. 13 pers, Publications and Proceedings. 43
Future Use Scenario .................................................. 13
6.2.Internet Sources .............................. 44
3.3. Functional Analysis ........................... 14
3.4.System Requirements .......................... 15
Functional Requirements ........................................... 15
Usability Requirements .............................................. 16
3.5.System Architecture ............................. 17 Appendix A - Feedback Selection ................... 45
System Modules ........................................................ 17
3.6.Conclusions ......................................... 18 Appendix B - GUI Concept Evaluations ........... 48
Heuristic Evaluation .................................................. 48
4. Detailed Design ......................... 19 Evaluations ............................................................... 51
4.1..................................................................MAppendix C - Sensor Specifications ................ 56
odules & Interactions.................................. 19 Accelerometer. .......................................................... 56

3
SensorShoe: Mobile Gait Analysis

Gyroscopic Sensors .................................................. 57

Appendix D - Interface Prototype .................... 58

4
 SensorShoe: Mobile Gait Analysis

^Introduction

l.l. Motivation
The main source of inspiration for this project is a similar project called GaitShoe1. In this project, a shoe
equipped with sensors can analyse the walking behaviour (gait) of a user, and provide the user with helpful
feedback. The focus of the project was with the development of the sensors and electronics, and the analysis
of sensor data. The sensors send their data through a wireless connection with a computer system, which
would translate the data to gait cues.

Fig. 1: The GaitShoe sensor-set, attached to a normal shoe

The report of the project recommends several useful applications for the GaitShoe, one of which is to allow
people to be monitored in a home-environment, unaffected by laboratory settings2. This is where the current
project continues.

1.2.Objectives
This project will look at the implementation possibilities of a GaitShoe for a home-environment. To be able to
design for a more specific target-group, some preliminary research has been done. Initial literature research
shows that Parkinson's Disease (PD) patients could benefit from the GaitShoe concept.
At the moment, PD patients with an early stage of PD are able to perform daily tasks, but are faced with their
reduced abilities everyday. Furthermore, though PD patients can live independently at home, they have to
pay regular visits to physical therapists for check-ups and exercises.
The goal of the current product concept, SensorShoe from now on, is to offer mobile real-time gait analysis
for PD patients, and remote monitoring and control by caregivers. The summarised objectives are listed
below.
1. Provide PD patients with an easy to use mobile gait analysis device
1. Enhance daily life of PD patients
2. Improve independence of PD patients
2. Provide caregivers with an effective remote monitoring and control tool for PD patients 1.
Save time of caregivers
1.3. Focus of this Study
The detailed design of the sensors on the shoe has already been described in the GaitShoe project, as
mentioned in the previous paragraph. This project will focus on the human-machine interaction between the

1
See http://www.media.mit.edu/resenv/GaitShoe/index.html
2
S.J. Morris, 2004

5
 SensorShoe: Mobile Gait Analysis

SensorShoe and the PD patient. Together with the detailed design of the shoe sensors this should result in a
feasible integrated SensorShoe concept.

Assumptions / Limitation
- We assume the SensorShoe to operate with accelerometers and gyroscopes:
- 3 axis of accelerometers; x,y,z
- 2 axis of gyroscopes; x,y
- The analysis and processing of raw sensor data is beyond the scope of this study

1.4. Approach
The following approach is used, based on the general approach for user interface design3.
1. Perform a small-scale literature research on Parkinson's Disease and Gait Analysis technology, to
explore these fields of science.
2. Use the results of the literature research to define a user and a use scenario
3. Determine the core system functions and requirements. In other words, describe what the system
should do and how the user interface should support this.
4. Develop interface concepts according to requirements.
5. Test interface concepts, by conducting heuristic evaluations.
3. Based on the results of the heuristic evaluations, one interface concept will be developed in to a
functional prototype, which may later be used for testing.

Result
The research will result in an interface design and prototype for the SensorShoe, aimed at PD patients.
Because of time limitations, the prototype will not be fully functional. Furthermore, usability testing will not be
conducted within this research. However, several recommendations can be given afterwards, with respect to
further development.
2. Literature Research
To introduce the context of the development of the SensorShoe, this chapter will start with an introduction
about Parkinson's Disease and PD patients. The effects of the disease will be described, followed by the
problems PD patients have to deal with.
Next, the possibilities of current gait analysis methods are investigated. The results of these investigations
provide a starting point for the rest of the SensorShoe development.

2.1. Parkinson's Disease

James Parkinson first described the disease in 1817 as "shaking palsy"4. It was only in 1960s that scientists
discovered that this disease was caused by a loss of brain cells which produce dopamine.
Parkinsonism was defined as the motor aspects of PD, this report will mainly focus on this part of PD.
- PD is most common with European people, it's much less common with Africans 5. East Asian people
have an intermediate risk. In rural areas there are more people with PD then in the urban area's.
Generally, more men are PD patients then women.
- In the USA one million6 people suffering from PD and about 40.000 thousand people are diagnosed with
PD each year. This is between 100 and 250 per 100.000 US citizens. In China this number is around

3
Stone et al, 2005
4
Langston, J.W., 2006
5
See http://en.Wikipedia.org/wiki/Parkinsons_disease, consulted on 22-3-2007
6
The Frequently-Asked Questions about Parkinson's Disease. New York, USA: Parkinson's Disease Foundation, 2006

6
 SensorShoe: Mobile Gait Analysis

1700 per 100.000 for those older then 65 years 7. In average 14.8 people per 100.000 in Finland have
PD.
- It's not common to have PD for people under 40 year, though it is possible. Generally, the symptoms are
beginning around the age of 58-60 (in the USA).The symptoms increase as the age increases, however,
around 15% of the PD patients are diagnosed before 50 years.

Causes
Parkinson's disease is a failing or braking down of brain cells which results in a loss of dopamine-
production3,5. The failing brain cells are located in the substantia nigra, which is an area in the brain.
Dopamine is a chemical which is a messenger for neurological signals; the signals mainly consist of
information for controlling the movement of a body.

Symptoms
Common symptoms of PD are8:
- Tremors of the limbs, jaw and face. The shaking of the body parts is uncontrollable [2], which makes
daily tasks more difficult. Even though this is the most well-known symptom, just 30% [3] of all PD
patients experience tremors.
- Rigidity, an abnormal stiffness of the muscles, of the limbs and trunk. This causes an increased risk
for falling.
- Bradykinesia. All movements become very slow; the walking speed decreases and one starts to
shuffle. This symptom is considered to be the most disabling of PD.
-
Postural instability causes PD patients to fall. The instability is caused by loss of balance and
coordination.
These symptoms may result in the following gait-related disorders:5
- Short steps
- Shuffling gait (festination)
- Freezing (difficulty initiating gait)
- Difficulty to turn
- Loss of balance
In general, it can be said that PD patients suffer of a decreased quality of life, since they are often not able to
perform daily tasks. Even small tasks can be difficult, such as getting out of bed.

Treatments

Medication
The purpose of medication for PD is bipartite.
First, medication can increase the levels of dopamine in the brains, secondly, the medication can help getting
the dopamine absorbed by the body. Levodopa (or L-dopa, a dopamine increasing medicine) is most
commonly used, 70 to 80% of the PD patients use Levodopa. Several other types of medication are used at
the same time to be able to reduce the symptoms of PD.
A medication which is used combined with Levodopa is Carbidopa; this is a medicine which is from the
second category. By using Carbidopa, more of the Levodopa is transported to the brain instead of being used
in the bloodstream. This results in a lower required dosage of Levodopa.

Surgery
Generally there are three kinds of surgery which are used to decrease the effects of PD. Firstly, Pallidotomy
which relieves rigidity and bradykinesia. Secondly, Thalamotomy that relieves tremors. In both procedures,

7
See http://en.Wikipedia.org/wiki/Parkinsons_disease, consulted on 22-3-2007
8
See www.pdf.org, consulted on 22-3-2007

7
SensorShoe: Mobile Gait Analysis

some regions of the brain are destroyed to decrease the symptoms. The third surgery method is to place
implants (electrodes) in the brain, which will generate electric impulses to decrease the symptoms.

Therapy
Regular exercise is important for PD patients. With exercise the muscles are trained. The muscle strength is
maintained and therefore the mobility, flexibility, balance, range of motion etc is maintained as well.
Besides regular exercise, different therapies exist. All these therapies help to control the symptoms of PD and
make daily life easier. Physical therapy helps, just like daily exercise, to maintain muscle strength. Speech
therapy helps increasing the voice volume and makes talking more easy. Occupational therapy helps PD
patients find other ways in doing their daily tasks which are difficult to perform.

8
SensorShoe: Mobile Gait Analysis

2.2.Gait Analysis
The above paragraph already mentioned problems with walking
caused by reduced muscle control. Physical therapy can be
used to teach the patient how to cope with these problems. At
home, patients can do regular exercises to keep their muscles in
shape. Furthermore, the effects of medication and therapy can
be evaluated by a physical therapist, using so called gait
analysis systems.
Gait analysis is primarily used as an objective and quantitative
method to record the effectiveness of physical therapy. it is a
method of recording the way people are walking in many
different ways. During the gait analysis, three characteristics are
measured: the temporal, kinetic and kinematic aspects of the
gait9.
- Temporal characteristics, often also seen as kinetic
aspects10, are the characteristics which often change
during therapy. They are easy to measure and should
often be measured. Examples of temporal
characteristics are stride length, cadence and velocity.
- Kinetic characteristics exist of gait which is caused by
external forces. These forces can be caused by the
ground or forces acting on the PD patient. Important is
the pressure carried out on the ground and thus the force on the knee.
- Kinematic characteristics include the movement of the Fl'9- 2: Test: pers°n in a current: gait analysis body itself,
important examples are the linear and laboratory angular displacements, velocities and acceleration.
The diagram below shows the basic structure of a gait analysis. The PD patient is being watched by the gait
analysis system (for example by video camera's or body sensors) while performing an exercise, such as
walking. By comparing the gait parameters of the PD patient with the parameters of a healthy age- matched
control person, the gait analysis results in a certain type of feedback.

The feedback could be 'increase your stride length' or 'change walking rhythm'. There are different ways of
sending this feedback to the PD patient. If the goal of the gait analysis is to evaluate the performance of a
patient, feedback is not given at all.

9
Melmick, M.E. et al. 2002
10
Peppe, A. et al. 2007

9
SensorShoe: Mobile Gait Analysis

2.3. Gait Feedback Cues

Several sources indicate that feedback cues could help the patient with improving their stride length, cadence
and other important gait parameters. Research has been done to determine which types of feedback can be
used, and how effective they are. This paragraph lists several of these feedback methods, along with their
supporting research and an indication of effectiveness.

Visual Cues
Several studies investigated the effect of visual cues on a patient's gait11, 12. A study of Martin13, it was found
that the stride length of PD patients could be increased by using visual cues perpendicular to the walking
path.
Continued research, for example that of Morris, further explored the possibilities of visual feedback. it is
shown that visual markers (SL markers) can increase the walking velocity of PD patients by increasing their
striding length. In the same experiment, the use of a subject mounted lighting device (SMLD) caused similar
improvements.

Baseline SL markers SMLD

Patient Control Patient Control Patient Control

n 14 14 14 14 13 14
Velocity (m/s) 1.06 ± 0.21 1.39 ± 0.22 1.17 ± 0.18 1.36 ±0.17 1.22 ± 0.19 1.41 ± 0.24
Cadence (steps/min) 120 ± 11.0 117 ± 8.0 105 ± 14.1 120 ±11.7 112 ± 11.7 113 ± 12.9
Stride length (m) 1.10 ± 0.25 1.42 ±0.18 1.34 ± 0.09 1.36 ± 0.07 1.29 ± 0.20 1.49 ± 0.1 7

Fig. 1: Results of the visual feedback experiment (From Morris, 2001)

Current research is working on an explanation for the effects of visual cues. No hard evidence is given yet,
but one theory says it may be due to the concious awareness of the
movements that's provided by the visual cues14. Being able to see the
movement could make it easier to actually carry out the movement. This
theory could also apply to the other feedback modalities.

While the above research primarily used directional cues (stripes),

another study used rhythmic visual cues to affect the gait rhythm of the
patient15. This form of feedback appears to be effective as well, but has
some drawbacks. Visual cues are disturbing for the environment of the
patient, and other sources of light may disturb the rhythm.
Another development in this area is the use of virtual reality. By
projecting objects on a virtual field of vision, a patient is able to improve
his/her gait. At the moment only short-term effects are found. Further
research should indicate the possibilities of Virtual Reality for long-term
effects.
Fig. 2: Test person equiped with a
SMLD

11
Lewis et al, 2000
12
Morris et al, 2001
13
Martin JP, 1967.
14
Morris et al, 2001
15
van Wegen et al. 2006

10
SensorShoe: Mobile Gait Analysis

Audio Cues
Auditory feedback can be used in several ways. First of all, users can be given spoken clues about their gait.
For example, a simple command of 'walk towards the door' can help the PD patient in fulfilling this task.
Spoken information about the walking speed and distance may also help.
A second way of using auditory feedback is to provide the user with rhythmic audio cues, or even music. As
walking is quite a rhythmic process, the presence of a stable rhythm is proven to be helpful for PD patients'
gait. An example of such auditory feedback can be found in the GaitShoe project16.
Auditory feedback can be provided through headphones, which prevent the user from annoying or disturbing
the environment. However, headphones may damage the ears, and don't allow the user to receive the audio
feedback while listening to others or while participating in traffic.

Somatosensory Cues
To solve the mentioned shortcomings of auditory feedback, some research has also been done in the field of
tactile or somatosensory feedback. This feedback uses the skin to communicate with the user, by applying
movement, deformations or pressure to the skin. This feedback can be used for rhythmic feedback, but may
also provide directional information (i.e. left/right directions).
Experiments have been conducted to investigate the relation between somatosensory feedback and possible
interference with visual feedback17. Feedback was given using an RSC (A vibrating device which was held by
the test persons) and visual feedback, by providing the test persons with a virtual corridor projected on a
screen while walking. The results are shown in the diagram below.

Fig. 3: Results of the somatosensory feedback experiment (From Wegen et al, 2006)

The results lead to the conclusion that somatosensory feedback has about the same effect as audio
feedback, and could therefore be seen as a feasible alternative. Furthermore, it is concluded that the
somatosensory feedback is not affected by the visual cues, which allows multi-modal feedback as well.
The major advantage of somatosensory feedback is that it can be sent to the user discretely, without anyone
noticing it. This aspect should be taken in to account during the further design of the SensorShoe.

1
6 Morris et al, 2001
1 van Wegen et al, 2006
7

11
SensorShoe: Mobile Gait Analysis

2.4. Conclusions
This chapter started with a review of Parkinson's Disease (PD). PD is a disease that causes failure of
braincells, resulting in a reduced dopamine production. In turn, this results in several symptoms, including
loss of limb control, stiffness and tremor of muscles. Treatments are available, but can not yet cure the
disease entirely. It was found that physical therapy plays an important role in controlling the disease and
reducing the effects.
The next part of the literature research investigated the possibilities of gait analysis methods for PD patients.
Several studies have already indicated that laboratory-based gait analysis can help PD patients by providing
them with feedback on their gait. The gait-related feedback, or gait cues, can be provided in different ways.
These have been further investigated and described.
Research shows that feedback cues, being visual, audio or tactile, can significantly improve the gait of PD
patients. An overview of advantages and disadvantages of these techniques has been presented, which
indicates that the functional use of each method is equal, but the usability of the methods differ.
Visual and audio feedback cues may annoy the environment of the user, while tactile feedback is usually
more discrete. This is an important consideration for the further design of the SensorShoe.

12
SensorShoe: Mobile Gait Analysis

3. Concept Development
The literature research indicates the need for the development of a SensorShoe for PD patients. This chapter
describes the development of the SensorShoe concept, which will later be further developed in the Detailed
Design chapter.
The diagram below shows the approach that will be used for the concept development. As shown, the first
step is to define the target group, including future users and future use of the product, described in a Future
Use Scenario.
From this scenario the main functions of the concept will be derived, describing what the concept should do.
The requirements will define the design space in which solutions for the functions have to be found. After
determining the functions and their requirements, an integrated system layout will be defined, describing the
system components and their interactions with each other.
This approach should result in a SensorShoe system concept suiting the needs of our target group.

13
SensorShoe: Mobile Gait Analysis

3.1. User Characteristics

The first step in the concept development phase is the definition of the target group. As said in the
introduction, the SensorShoe will be designed for PD patients. However, the PD patient will not be the only
end-user of the product.
The caregiver is another important stakeholder. He/she will use the SensorShoe to remotely monitor and help
the PD patient. Therefore it is important to also consider the characteristics of the caregiver in this design.
Each of these users is described by the following user characteristics, which are mostly based on information
from the literature research of chapter 2.

PD Patient
- Age: mostly 50 and up
- Sex: both male and female
- Motor skills: Limited18
- Slow movements
- Uncontrollable movements
- Limited reach of limbs
- Cognitive skills: No specific PD-related limitations, but
- Decreasing sight and hearing due to ageing
- Decreasing memory capabilities due to ageing
- Computer experience: limited or none

Caregiver
- Age: 30 to 60
- Sex: both male and female
- Cognitive skills:
- Assumed to have no cognitive disabilities
- Medical knowledge
- Physical knowledge
- Computer experience: daily use of administrative software and home-use
- Limited mobility
- Supervises large patient group
3.2. Use Scenario
it is common for a product design project to ask users what they'd like to see in a product. However, in case
of designing new/future products, it is hard for people to imagine what they could do with, or want from such a
product. As an alternative to questioning people, it is therefore decided to use a Future Use Scenario.
The scenario describes a future in which the product already exists. The characteristics of the previous
paragraph can be used to create persona's, which can be placed in to the future use scenario. Imagining the
persona's in the future use scenario results in a product description that can be used for further development.
To do so, two persona's will be created first. They represent a future PD patient and a future caregiver. After
they've been created they will be placed in to a Future Use Scenario.

18
See paragraph 2.1

14
SensorShoe: Mobile Gait Analysis

Persona's
PD patient - Dirk is a 53 year old man, who lives on his own in a small house. Dirk has been a PD patient for
about 10 years now, which causes Dirk to have problems with daily tasks. Most effected are the muscles of
his legs, so walking around the house is quite difficult.
Caregiver - Mary is 32 year old physical therapist. She runs a clinic in a medium sized town, where she
supervises about 20 PD patients. Dirk is one of these patients. She has regular appointments with the PD
patients, but needs time to help other patients as well.

Future Use Scenario

When he awakes, Dirk is already faced with PD. it is hard to get out of bed because his muscles are stiff, and
he hasn't taken medication yet So Dirk stumbles towards the bathroom, takes a shower and gets his clothes
on. With that, Dirk also puts on his pair of SensorShoes.
Without consciously knowing it, Dirk's gait is being watched and analysed by the SensorShoe. The data is
logged and later sent to Dirk's physical therapist.
After breakfast, Dirk decides to do some shopping. The walk to the supermarket used to take him about 5
minutes, but this time increased ever since he got PD. Before Dirk leaves the house, he 'activates' the
SensorShoe. This means that the SensorShoe will provide Dirk with real-time feedback on his gait, improving
speed and stability.
While the first 10 meters are quite unstable, Dirk notices that by following the feedback cues of the
SensorShoe he improves the stability of his gait. After a while he uses the feedback cues on an unconscious
attentional level.
In the afternoon, when all the shopping is done, the SensorShoe reminds Dirk of his daily exercises. These
exercises help Dirk keeping his muscles in shape, and slow down further PD-related symptoms. The
SensorShoe interface informs Dirk about which exercises he should do, and for how long. After finishing the
exercises, the results are sent to Dirk's physical therapist, Mary.
Mary receives the daily or weekly log of Dirk's activities. The log indicates which activities have been
detected, and how Dirk was supported by cues. For example, the log contains a report of the 15 minutes walk
to the supermarket, also reporting the feedback cues that were used. Additionally, the therapist can analyse
the effects of the daily exercises on the performance of Dirk.
After analysing these data, Mary can adjust the training programme or feedback cues to optimise the effects
for Dirk.

15
SensorShoe: Mobile Gait Analysis

3.3. Functional Analysis

Based on the scenario, a set of functions has been defined which describe the functional behaviour of the
system concept. The functions describe what the system should do, but do not yet define how the system
should do this.
- Analyse Gait - The main function of the SensorShoe is to measure and analyse the gait of a person with
Parkinson's Disease. The gait will be analysed based on the acceleration and gyroscopic changes of the
feet. After analysing the raw data three entities should be known: walking speed, cadence and stride
length.
- Provide Real Time Gait Feedback - The SensorShoe should be able to provide the user with real time
feedback on his/her gait. This means that the user should be able to walk with the SensorShoe, while
receiving cues on gait improvements.
- Provide Physical Therapy and Feedback - The caregiver can make a selection of several exercises
which the patient has to do daily. This is the physical therapy which is done at home. The user can
receive feedback on these exercises.
- Evaluate the Feedback - For a caregiver, it's important to know how a patients experiences the
exercises. Based on these evaluations the caregiver can alter the physical therapy exercises to fit the
needs and preferences of the patient.
- Record and send Log to Caregivers - To know how a patient progresses during the physical therapy, the
SensorShoe will record the gait of a patient and send it to the caregiver.
- Receive PT Settings from Caregivers - The caregiver receives the information (evaluation and gait
analysis) at his office. On his own computer he can interpret the results and adjust the physical therapy.
These adjustment need to be send to the SensorShoe.

16
SensorShoe: Mobile Gait Analysis

3.4.System Requirements
As said, the functions only describe the functional behaviour of the system. Technical solutions have to be
found for each of these functions, and in the end be integrated in to a final product. To define the design
space of the functions, system requirements will be used. Requirements indicate limitations and conditions
for the functions.

Functional Requirements
Functional requirements define the design space for the functional solutions of the concept. The list below
presents requirements for each of the main functions of the SensorShoe.

Analyse Gait
- The SensorShoe must measure the walking speed (m/min)
- The SensorShoe must measure the cadence (steps/min)
- The SensorShoe must measure the stride length (m)
- The SensorShoe must measure the vertical displacement of the foot (mm)

Provide Real Time Gait Feedback

- If desirable the SensorShoe will provide feedback to the user during a walk, this to support the cadence
and stride length.

- The user should always be able to turn off and turn on the feedback

Provide Physical Therapy & Feedback

- The SensorShoe helps to remember the user to its daily physical therapy
- The SensorShoe must accompany the user during the physical therapy. (indicate what to do and the
amount; when an exercise is not performed right, the SensorShoe should give suggestion how to
improve the performance)

- The user should always be able to turn off and turn on the feedback

Evaluate Feedback
- The user should be able to influence the behaviour of the feedback of the SensorShoe
- After the exercise an evaluation can be given by the user. Whether the exercises weren't to fast, to many,
to difficult or to easy, to slow or to little. Based on this information, the caregiver can adjust the Physical
Therapy.

Record and send Log to Caregivers

- The SensorShoe must measure the progress in every exercise (both PT and walking support)
- The SensorShoe should store the gait data in order to be able to analyse it (pre analyse raw data)
- The SensorShoe must be able to send the raw, and pre analysed, data to the caregiver
- The SensorShoe should also be able to send the users evaluation to the caregiver

17
SensorShoe: Mobile Gait Analysis

Receive PT Settings from Caregivers

- The caregiver must be able to interact with the SensorShoe from a distance
- The caregiver can change the settings of the SensorShoe. The daily Physical Therapy exercises can be
adjusted as a result of the gait analysis. The caregiver can change the Physical Therapy content or the
amount of it.
- Also, the caregiver must be able to send short text messages to the user, in order to explain the changes
in the PT content.

Usability Requirements
Usability requirements define how the user should (be able to) operate the SensorShoe. The following
requirements are quite generic, and will be further specified during the detailed design of the
SensorShoe.

Interface Usability
- Input and output of the SensorShoe interface should support the capabilities and limitations of the user
group
- Screen output should be readable for our target group
- Audio output should be receivable for our target group
- Tactile input should be comfortable

Practical Use
- Since the device is portable, one should be able to carry the device around.
- The concept should use a (rechargeable) mobile power supply
- The concept should be as lightweight as possible
- The concept should be compact
- The user should not feel stigmatised while operating the concept

18
SensorShoe: Mobile Gait Analysis

3.5.System Architecture
With the desired functionality and requirements defined, a system concept can be made. The system concept
describes a global system architecture. This architecture should support the functionality and requirements.

System Modules
The global architecture of the system consists of three modules. Each of these modules takes care of one or
more functions defined earlier. The following list shows the allocation of functions over the system modules.
-Feedback -Processing -Gait sensors
-
Communication

-User Interface
- Sensor-equipped shoe - This shoe is being developed by another project group and will therefore not
be described in detail here.
- Analyse Gait
- Interface unit - The project group prescribes a PDA to be used. The interface for the PDA will be
designed and described in detail in the following paragraphs.
- Provide Physical Therapy and Feedback
- Evaluate the Feedback
- Record and send Log to Caregivers
- Receive PT Settings from Caregivers
- Feedback unit - The third part of the architecture consists of a feedback unit. The feedback unit will send
the gait cues to the user. As found in the literature research, several types of feedback can be used,
including auditory, visual and tactile19. These possibilities will be evaluated in the next paragraph, after
which a selection can be made that is to be used within this design.
- Provide Real Time Gait Feedback

19
See paragraph 1.4

19
SensorShoe: Mobile Gait Analysis

3.6. Conclusions
Chapter 3 started with describing the future users of the SensorShoe. Both the patient and the caregiver have
been described and were placed in a future use scenario. From the scenario the major functionalities and
requirements have been derived. The main functions of the SensorShoe are listed below.
- Analyse Gait
- Provide Real Time Gait Feedback
- Provide Physical Therapy and Feedback
- Evaluate the Feedback
- Record and send Log to Caregivers
- Receive PT Settings from Caregivers
After defining the functions a system architecture has been designed, which should include all of the
SensorShoe functions. The architecture consists of three modules, namely the shoe, a feedback device and
a PDA. All functions have been allocated to a particular module.
The next chapter will further describe the detailed specifications of these modules, as well as the interactions
between the modules.

20
SensorShoe: Mobile Gait Analysis

4. Detailed Design
The previous chapter described the global architecture of the SensorShoe concept. This resulted in a system
layout containing three modules, namely the Shoe, the PDA and the Tactile Feedback component. In this
chapter each of these modules will be further specified, and integrated in to a final detailed design.

4.1. Modules & Interactions

The interactions between the modules (exchange of information) are to be specified. A data-flow diagram of
the entire system has been made, as shown in the diagram below.
The diagram contains the three modules, indicated by the coloured rectangles. Each module contains
functions and data-sets which support the required functionality for each module. For example, the shoe
contains the data-set "Current Gait" which provides the PDA with information about the gait of the user.
The diagram also indicates how the modules interact with the user and/or caregiver. The dotted horizontal
lines indicate the border between user, caregiver and 'system'. These interactions can be used to design the
user interface for the PDA.
The internal functions and data-flows of the system have been colour-coded:
- Green arrows represent data exchange between the PDA and the user
- Blue arrows represent data exchange between the Tactile feedback unit and the user
- Red arrows represent data exchange between the Shoe and the PDA
- Purple arrows represent data exchange between the caregiver and the PDA
- Black arrows represent PDA internal data flow
The data flow diagram can be used for the further implementation of the PDA software. In this study it will be
mainly used to specify the interactions between modules, and the interactions between the user and the
system.

21
 SensorShoe: Mobile Gait Analysis
User Requests PT PT Evaluation User preferences on User Requests real
real time feedback time feedback
PT Feedback k Gait
feedback

Exercise Log

Evaluate Exercises
n+1, next exercise Finished exercise

Loop: next exercise User Preferences

Initiate Physical Therapy Select Exercise Compare Gait Initiate real time Gait
Analysis
Start realtime
nth bxerdse
analysis

Available
exercises Req.Gait
Track Timing
PT Exercises &
Required Gait
Time for
exercises

PT Changes

Start Log Start Log

Gait Log

Communication Update PT Exercises

--Caregiver

C aregiver receives C aregrver ch an ges

log PT settings
SensorShoe: Mobile Gait Analysis

4.2.Shoe Module
Though the shoe and its electronics is being developed by another project group, the required functionality
and specifications will be presented here. This should ensure compatibility between the user interface and
the shoe.

Sensors
The main function of the shoe is "Analyse Gait"20. As mentioned in the first chapter, the Shoe component will
use several types of sensors to measure current gait parameters. These sensors include:
- x,y and z accelerometers
- x,y gyroscopic sensors
Comparable accelerometers an gyroscopic sensors have been used in the GaitShoe project. The
specifications of those sensors have been used for the further design of the SensorShoe as well. These
specifications are available in appendix C.
The sensors will start measuring as soon as the shoe is put on. The main reason for this is that more gait-
data results in a more reliable analysis. Furthermore, the shoe will not perform the analysis itself. This would
require too much (processing) power. Therefore the data will be sent on a constant interval to the PDA21. The
required communication for this function is discussed in the next section.
Furthermore the sensors should be power efficient, so that the SensorShoe may be used for a day (24 hours)
without recharging. Because the exact sensor specifications are not available for the SensorShoe project yet,
it is hard to give precise information about the battery requirements. However, sources indicate that a
Lithium-ion Polymer (Li-Polymer) battery provides sufficient power while also offering design flexibility22.
The shoe module will be equipped with the following components.
Component Power usage (peak/avg, mW) Dimensions

Accelerometers 23 unknown @ 3V 4 mm x 4 mm x 1.5 mm, < 1,0

ADXL-330 gr

Gyroscopes24 unknown @ 3V 5 mm x 5 mm x 1.5 mm,

IDG-300 2D Gyroscope < ~1,0 gr

Bluetooth module 90 / 0.09 mW -

Communication
To send the sensor output to the PDA, the SensorShoe will use a class 2 Bluetooth connection. Bluetooth
offers several advantages25.
- Low power consumption
- Low-cost chips
- Small and light-weight hardware
- Sufficient bandwidth and range

20
21 See Chapter 3
22 We assume the analogue data from these sensors to be converted to a signal which can be sent by the the Shoe to the PDA.
23 See http://en.wikipedia.org/wiki/Lithium_polymer
24 See Appendix C
25 See Appendix C
See http://www.silicon-press.com/briefs/brief.bluetooth/index.html

23
SensorShoe: Mobile Gait Analysis

Bluetooth has been designed for low-range and low-energy connections. Furthermore, Bluetooth is a
common communication method on modern PDA's. Class 2 Bluetooth supports a maximum connection range
of 10 metres, with a maximum permitted power of 2.5 mW (4dBm). This class has been chosen because of
the range it offers, which should be at least 1 metre (distance between the foot and hand- height).

4.3. PDA Module

The PDA is a major part of this concept. it is decided to use a PDA because of it is relatively cheap, and
available for consumers. Furthermore, a PDA offers a development platform for which a scalable user
interface can be designed. This paragraph will present detailed specifications and requirements for the PDA,
as well as the design of the GUI for this device.

Specifications
The following specifications should be met by the PDA in order to work with the SensorShoe.
- Bluetooth connection, which will be used for communication with the Shoe component and the Tactile
Feedback unit.
- Long-life battery, which should allow the PDA to operate (with Bluetooth enabled) for about 24 hours
without recharging.
- A visual display with touch screen, with a minimum screen size of 3,5", which means 71.1 mm x 53.3
mm, and a minimum resolution of 240x320 pixels.
- Support for audio feedback
As an example, we've used the HP iPaq hx2490 PDA26, which meets these specifications.

Functionality
As defined in chapter 3, the PDA will take care of several important system functions. These functions will be
provided to the user trough the graphical user interface of the PDA.
- Provide Physical Therapy and Feedback
- Evaluate the Feedback
- Record and send Log to Caregivers
- Receive PT Settings from Caregivers
The first two functions are most relevant for the patient, while the last two functions are only useful for the
caregiver. The focus of this project is on the design of the PDA GUI for the patient.

GUI Structure
The design of the GUI starts with defining the structure of the interface. This structure is shown in the
Decision Diagram below. The diagram starts with the 'main screen' of the interface, from which the user can
choose an option. These options provide access to the main functions of the PDA:
1. Start Physical Therapy Session
2. Start Real Time Gait Analysis
3. View Patient Progression
4. Check System Status
Each option provides the user with new choices, as presented in the diagram. This diagram can be used to
see which items should be available in which parts of the interface.

26
See http://www.pdashop.nI/product/19068/hp-ipaq-hx2490.html#specificaties

24
SensorShoe: Mobile Gait Analysis
KJ
UJ
SensorShoe: Mobile Gait Analysis

Interface Concepts
Having defined the GUI structure, the actual interface screens can be designed. These design concepts
consists of some sample screens, presenting a concept layout and widgets. The concepts differ in user
approach, usability and 'look and feel'. After presenting these concepts, a heuristic evaluation will be used to
evaluate the concepts, and decide which of them is feasible enough to further develop.

Concept 1
Concept 1 is the most 'standard' design of all five. The header tells the user which function is used (Physical
Therapy, Real-time gait etc.). The footer gives the user extra information:
- how many exercises still to go?
- How many repetitions?
The three buttons on the left are the choices the user has: i.e. pause exercise, help or next. With Realtime
gait the user can increase or decrease the feedback of the walking speed. Different background colours

indicate different functions.

26
SensorShoe: Mobile Gait Analysis

Physical Therapy

w KJ
Exercise 1/5
Still 6 repetitions

27
SensorShoe: Mobile Gait Analysis

Concept 2
The second concept is a less conservative concept. This concept is based on a flower. For a flower it is
important to attract bees to the core of the flower. The most important part is played by the leaves, which lure
the bees to the flower.
In this interface flower leaves play an important role as well. The leaves hold the main menu item, and
organize them in a natural sense. Each leave contributes in order for the core to be at its best. That's not only
in nature, but also appropriate for the interface. Without the menu leaves the core would be useless.
Four circles indicate the four functions. When the user presses a circle, the function is selected. The other
three circles move to the background. The selected function blooms: the buttons appear.

Concept 3
In concept 3 the header is placed on the right. The user can see which function is activated. At the bottom of
the interface, three buttons, with triangle shapes on top, are located. Each of them expand when pressed.
When expanded, it provides the user extra information, or extra buttons, which are also described in concept
1.

28
SensorShoe: Mobile Gait Analysis

Concept 4
Concept 4 uses a drop down menu. When the menu button is pressed, it drops down over the current used
function. This drop down menu provides extra information as well as extra buttons.

Concept 5
This concept is based on a calendar. The user chooses a week and a day. This day contains several
exercises which are explain in the next screen. Located at the bottom are several buttons. The header tells
the users which function is currently activated.

29
SensorShoe: Mobile Gait Analysis

Concept Evaluation & Selection

The concepts presented above have been evaluated using existing guidelines for GUI design27. These
guidelines apply to the design for disabled users, general GUI's and website-design. Irrelevant guidelines
have been left out. Each concept scored a certain amount of points for each guidelines, resulting in the total
scores presented in the table below28.

Concept 1 Concept 2 Concept 3 Concept 4 Concept 5

2nd 2nd 2nd 2nd 2nd
1st
1 st
1st
1st
1st

Guidelines GUI 23 13 30 9 12 14 23 8 17 14
Guidelines website 6 2 10 2 4 3 7 3 9 2
Guideline disability 7 0 8 0 3 0 7 0 5 0
Sub total 36 15 48 11 19 17 37 11 31 16
Inc. potential 51 59 36 48 47

Based on the scores, it is concluded that concept 2 meets most guidelines and offers the best potential to
improve current design flaws. Concept 2 is further developed using a functional prototype, which in turn will
be used for a usability test. The results of this test should indicate whether the concept interface works or not,
and which aspects need further improvements.

27
See paragraph 1 of Appendix B for an overview of guidelines used
28
Detailed evaluation scores have been included in paragraph 2 of Appendix B

30
SensorShoe: Mobile Gait Analysis

Detailed Interface Design

In the previous paragraph the concept designs were evaluated according to the guidelines. Since each
concept had its positive features and the concept design is not fully worked out, the positive features of both
will be combined. The following sections describe some of the major parts of the final interface. Other parts
are shown in Appendix D.
Main Screen

The picture on the left in the above table shows the main screen. The launchpad contains of five buttons, of
which four are the main features. The large buttons are the most important features, Physical Therapy and
Real time gait support, and the smaller buttons are the more advanced features, system status and
progression information. The fifth is the shut-down button. The colours and styling have been chosen in order
to create a playful yet serious interface.

Real Time Gait Support

The above screen on the right represents the real time gait support. In this
screen one can see the speed of the tactile feedback. In the below part, the
arrow points to the current speed. The care giver can pre-set a speed, which
is represented by the largest line below. The user can adjust the speed with
the "+" and the "-" symbol. The arrow will move as well. When one wants a
break, he can press the stop (the square) button, and the tactile feedback
will stop.

Physical Therapy
The Physical Therapy has five main buttons. These are located around the
picture of the exercise. These are: forward (the following exercise), stop (the
pause the exercise), help (the question mark icon, this provides an audio
support, as well as an animation of the exercise), "1/2" (one out of two, this
tells the user which exercise he is performing) and "9" (this tell the user how
many repetitions he still has to do).

31
SensorShoe: Mobile Gait Analysis

4.4. Feedback Module

The main function of the feedback unit is to provide the patient with Real-Time feedback on his gait. For the
detailed design of this module, the following questions are to be answered.
- Which feedback modality should be used?
- How can the modality be implemented in the SensorShoe design?

Feedback Requirements
The literature research indicated no major drawbacks or advantages for any of the feedback methods 29. To
select one of the modalities however, a set requirements for the modality has been made.
1. The modality should provide effective feedback
2. The modality should not be stigmatising
3. The modality should be implementable in the SensorShoe architecture;
1. The modality should be wearable
2. The modality should be power efficient
Within these limits, available feedback modalities have been listed and evaluated. These are presented in
Appendix A. Pro's and con's of each implementation have been included. After presenting the possible
solutions, a selection can be made.
For this selection, it was found most important for the user to be able to receive feedback in a non- intrusive
way. The visual feedback method doesn't meet this requirement, as it sends out visual cues to the user and
his/her environment. Auditory cues trough earphones are more discreet, but prevent the user from interacting
with the environment. These considerations lead to the decision to use tactile feedback.

Tactile Feedback
T
a
ct
il
e
fe
e
d
b
a
c
k
c
a
Fig. 4: Schematic representation of the skin n
b
e provided without non-users noticing it, and several
studies have proven the effectiveness of this method.
The detailed design of a tactile feedback unit will be
presented in the next chapter.
Tactile feedback relies on the ability of the user to
sense tactile signals (for example vibrations) trough
their skin. As shown in the diagram 30, the skin consists
of two layers (the Dermis and the Epidermis), and
contains several types of nerve endings:
- Thermoreceptors sense temperatures and changes
of temperature.

29
See Chapter 2 for an overview of feedback methods
30
T. Jyrinki, "Perception of Small Device Vibration Characteristics - Test Facilities Setup and Study"

32
SensorShoe: Mobile Gait Analysis

- Noiceptors sense damaging effects, such as intense

heat or high pressure.
- Mechanoreceptors sense indentations (pressings), and can be divided in to slowly adapting
mechanoreceptors (SA) and rapidly adapting mechanoreceptors (RA). Merkel's disks and Ruffini endings
are examples of SA mechanoreceptors. Meissner's corpuscles and Pacinian corpuscles are RA
mechanoreceptors.
Tactile feedback systems stimulate one or more of these nerves.

33
SensorShoe: Mobile Gait Analysis

For example, electro-tactile displays use small electrodes placed on the skin to send electronic signals. The
effectiveness of electronic signals depends on the humidity of the skin, the affected area and the applied
pressure. This is why it is hard for electro-tactile techniques to become commercially attractive; they're very
dependant on individual preferences.
An alternative to electro-tactile techniques are called 'mechanical actuators'. Mechanical actuators physically
affect the skin, either by higher frequency (vibration) or lower frequency (pressure, deformations) signals.
- Vibrations should be below 1000 Hz, which is the sensitivity limit for the human skin. A sensitivity peak is
found around 250 Hz. Besides the frequency, the sensitivity for vibrations is also affected by the density
of nerves in the skin. The skin in the face, tongue and hands is most sensitive31.
- Lower frequency signals should be within 20-40 kPa32, which is the range in which pressure is best
perceived by the user.
For the SensorShoe it is decided not to use electro-tactile feedback because of the drawbacks mentioned. As
the SensorShoe will be used by patients at home we do not want to run the risk of letting these these users
hurt themselves because of wrong settings.
Therefore, the SensorShoe will use mechanical actuators to provide the user with rhythmic cues as gait
assistance.

Mechanical Actuators
Several technologies are available to enable mechanical tactile feedback. The following table showsthe
advantages and disadvantages of the available techniques33.

Technology Description Advantages Disadvantages

Electromagnetic Electromagnetic motors - Easy to control
produce torque with two time- - Heavy components
Motors - Clean, quiet
varying magnetic fields,
caused by two coils or a coil - Easy design and - Low power densities at
and a magnet installation small scales
- Heat dissipation
problems
- Low static force capability

Hydraulics - High mass

- Force capability,
A hydraulic fluid is pressurized - Tendency for fluid leaks
by a power plant, controlled by - power output,
servovalves and delivered to - Design difficulty
rotary or linear actuators - stiffness, and
- Expensive
through pressurized fluid lines - bandwidth unmatched by
other technologies

Pneumatics - Relatively low bandwidth

- Good static force
A gas (normally air) is capability - Low actuation stiffness
pressurized by a power plant, - Low power capability
- Lighter than hydraulics
controlled by servovalves, and
delivered to rotary or linear - Pneumatic power plants
actuators through pressurized and distribution systems
fluid (air) lines easier to manage than
hydraulics

31
T. Jyrinki, 2004
32
See http://www.e-skin.ch
33
Hasser, 1995, 1996

34
SensorShoe: Mobile Gait Analysis

Technology Description Advantages Disadvantages

Piezoelectric -High forces at low speeds
Piezoelectric motors translate in small package - Requires precision
the vibration of piezoelectric machining
materials to linear or rotary - Very small components
- Necessary power gating
motion using frictional forces to can cause annoying and
produce usable torques or potentially hazardous
forces at low speeds, without noise, depending on the
the need for gear reduction. design
Magneto Magneto restrictive materials - High forces at low speeds
change shape when subjected in small package - Necessary power gating
restrictive
to magnetic fields. Magneto can cause annoying and
restrictive motors also potentially hazardous
mechanically rectify small noise, depending on the
oscillatory motions of the design
driving element(s). - Heat dissipation can be a
problem
- Requires precision
machining

Shape Memory SMA wires and springs - Good power-to-mass - Heat dissipation
Alloy contract when heated and ratio problems limit relaxation
expand again as they cool rate of wires
under stress. - Low efficiency during
contraction - Limited bandwidth

Practical requirements such as lightweight and portability cause the pneumatic and hydraulic solutions to be
not-feasible. Using a gas or fluid in a device like the SensorShoe also increases the chance of system
failures, and require more maintenance.
The remaining technologies have comparable advantages and disadvantages.
Piezo actuators are already being used in PDA's and other computer-related hardware to enable tactile
feedback. These actuators are able to produce vibrations within 200 to 250 Hz, using less power compared to
electro mechanical motors.
Similar projects, such as "Tactile Vest"34 and "ActiveBelt"35, use electro mechanical motors for tactile
feedback. Though the vibration frequencies are outside the optimum scope of human skin, they've been
reported to work for tactile-feedback tasks.
Primarily based on the reports of successful use of electro mechanical motors, it is decided to use these
actuators in the SensorShoe as well.

34
L. Jones et al
35
K. Tsukada et al,2004

35
SensorShoe: Mobile Gait Analysis

Actuator Implementation
One of the major advantages of using electromagnetic motors for tactile feedback is that these motors are
small enough to be integrated in to existing products. This should prevent a stigmatising effect of the
SensorShoe; other people can't see whether someone is using or wearing the SensorShoe components.
Several possible integrations into existing products have been investigated. They are presented and
discussed in the table below.

Ring Pro's Con's

The device can be in any The technique (for taps and communication)
shape, so the personal should be very small to be able to fit into a
likings can be a factor in ring. When this is even possible, it will be
choosing the product. very expensive.

The fact that it is possible in any shape

makes it also difficult to produce. The
product show be individually changed,
which makes it more expensive to product.

Belt Pro's Con's

No one notices that the user Since the belt is no directly placed on the
is wearing an extended belt, body of the user, the pants will reduce the
since no one can even see signal strength. Therefore it can be
the belt. possible that the user will not notice the
taps correctly.

Pro' Con'
s
Almost everybody wears a s every user, a watch should be made
For
watch, so no one will notice with a different look and feel. Personal taste
the user wearing a rhythmic is of great deal. This diversity causes a
tap device. more expensive product.

36
SensorShoe: Mobile Gait Analysis

Bracelet Pro's Con's

The bracelet can be hidden
under the clothes. As can a
watch, but this bracelet does
not need to replace the watch
and does therefore not need to
meet personal preferences
(functions etc)

A arm/wrist is the best location

to use tactile feedback, it's the
most sensitive place.

Clothes Pro's Con's

The actuators can be used In order to maintain the right strength of the
wherever a user is most feedback, the clothes need to be tight. This
sensitive. The user will not can be very unconvertible for the user (i.e.
miss any cues. hot summer days)
The clothes can be made in Also with this solution, the product needs to
any shape or worn under change for every user. I.e. everybody has
another set of clothes, in order another size.
to hide the technology.

Hat Pro's

Con's
Since the rhythmic tap Not everybody likes to wear a hat, this
device is inside the hat, no might be very annoying for the user. one
will know that the user is wearing a special hat.
The actuator signal may lose strength
since the signal needs to pass the hair.

37
SensorShoe: Mobile Gait Analysis

After reviewing the concept solutions, it is decided to use the bracelet as a platform for the tactile feedback
unit. The advantages of this solution are:
- A bracelet is small, and can be worn under clothes. They can easily be designed for different users, for
example male/female or young/old
- No adaptations to or additions on existing clothes or accessories are needed
- A bracelet offers sufficient space for a power source and mechanical actuators
- A bracelet is worn on a sensitive part of the body (lower arms)

38
SensorShoe: Mobile Gait Analysis

4.5. Conclusions
The above paragraphs further specified the components of the SensorShoe concept. The shoe, the PDA and
the tactile feedback unit have been fully defined, and result in the total SensorShoe concept.
The shoe component has been only been specified in terms of required functionality and specifications. The
main reason for this is that the shoe is being developed by another project team. The design of sensors and
data analysis methods is beyond the scope of this research.
The PDA component has been specified in terms of required specifications. A specific type of PDA is
proposed, which will also be used for further interface testing. A GUI for this PDA has also been designed.
Several concepts have been presented, after which one of them is selected for further development and
testing, based on a heuristic evaluations.
Required specifications and technology for the tactile feedback unit have also been investigated, which lead
to the use of a micro-vibration mechanical actuator. The mechanical actuator can be implemented in a
bracelet.
As a result, the SensorShoe offers the user a practical and usable tool for real-time gait analysis and physical
therapy assistance. The components use available technology, and have been designed in such a way that
they can be used discretely.
5. Evaluation & Recommendations
This report focussed on the user-interface of the SensorShoe project. Other components, such as the
sensors for the shoe, are being developed by other project groups. To support integration of the different sub
projects, this chapter will look at future work and recommendations for the further development and
implementation of the user interface.

39
SensorShoe: Mobile Gait Analysis

5.1. Evaluation
A first short evaluation of the presented SensorShoe concept is done using the system requirements. The
following table shows a review of the requirements. The 'fulfilled' column indicates the status of a
requirements, which can be 'x' for fulfilled, '-' for not-fulfilled or 'o' for 'fulfilled, but room for
improvement'.

Requirement Fulfilled? Remarks

Analyse Gait
The SensorShoe must measure the
x
walking speed (m/min)

The SensorShoe must measure the

x
cadence (steps/min)

The SensorShoe must measure the

x
stride length (m)
Provide Real Time Gait Feedback

x
If desirable the SensorShoe will
provide feedback to the user during
a walk.

The user should always be able to

x
turn off and turn on the feedback
Provide Physical Therapy &
Feedback

x
The SensorShoe helps to
remember the user to its daily
physical therapy
o The understandability of the physical therapy feedback
The SensorShoe must accompany
should be tested
the user during the physical
therapy.

The user should always be able to

x
turn off and turn on the feedback

40
SensorShoe: Mobile Gait Analysis

Requirement Fulfilled? Remarks

Evaluate Feedback
x
The user should be able to
influence the behaviour of the
feedback of the SensorShoe

After the exercise an evaluation can

x
be given by the user.

Record and send Log to

Caregivers
x
The SensorShoe must measure the
progress in every exercise (both PT
and walking support)
o
The SensorShoe should store the
Further development of the analysis software is required
gait data in order to be able to
for implementation
analyse it (pre analyse raw data)
o
The SensorShoe must be able to
The communication between the PDA and the caregiver
send the raw, and pre analysed,
has not been designed in detail yet
data to the caregiver
o
The SensorShoe should also be Option is available in the GUI, but communication between
able to send the users evaluation to the PDA and the caregiver has not been designed in detail
the caregiver yet

Receive PT Settings from

Caregivers
o
The caregiver must be able to
The communication between the PDA and the caregiver
interact with the SensorShoe from a
has not been designed in detail yet
distance
o
The caregiver can change the Option is in the GUI, but the communication between the
settings of the SensorShoe. PDA and the caregiver has not been designed in detail yet

The caregiver must be able to send o

short text messages to the user, in Option is in the GUI, but the communication between the
order to explain the changes in the PDA and the caregiver has not been designed in detail yet.
PT content.
The caregiver should also be provided with a special GUI
for sending messages, PT changes and analysing patient
performance.

41
SensorShoe: Mobile Gait Analysis

Requirement Fulfilled? Remarks

Interface Usability

Screen output should be readable GUI design is based on guidelines, but requires a usability
o
for our target group test for more reliable conclusions

Audio output should be receivable GUI design is based on guidelines, but requires a usability
o
for our target group test for more reliable conclusions

Tactile input should be comfortable o Requires a usability test for more reliable conclusions

Practical Use
o
More details on electronic components can be used to get
The concept should be as
a more reliable estimation of concept weight ans power
lightweight as possible
usage
o
More details on electronic components can be used to get
The concept should use a
a more reliable estimation of concept weight ans power
(rechargeable) mobile power supply
usage

The concept should be compact

o Requires a usability test for more reliable conclusions
The user should not feel
stigmatised while operating the
concept

The requirements-based evaluation is just a first step towards the full evaluation of the SensorShoe concept.
Nevertheless, several important recommendations for further development of both the interface part and the
general system concept can be derived:

System Concept Development

1. Communication between the PDA and the caregiver should be further investigated. The PDA may
use Wifi to connect to a home-network with Internet access.
2. The PDA software needs further development, based on the structure presented in Chapter 4. This
requires more information about the sensor data and communication protocols (Bluetooth and Wifi
for example).
3. The specifications of electronic components of the Bracelet and SensorShoe need to be investigated.
This will result in a more reliable estimation of final weight and dimensions.

Interface Development
1. A usability test should be conducted to provide more insight into the effectiveness of the GUI. The
focus should be with the interpretation of real time gait feedback and the physical therapy cues.

The next paragraph will present a more detailed approach for the further development of the user interface,
based on a usability test.
5.2. Usability Test
Though the GUI for the PDA has been designed according to GUI design guidelines, there's no guarantee
that the interface is as effective as expected. To verify the effectiveness of the interface, a usability test can
be conducted. The test may indicate flaws in the current design, as well as recommendations and feedback
from actual users.

42
SensorShoe: Mobile Gait Analysis

Based on the evaluation of requirements, the following aspects are relevant for further investigation in a
usability test.
1. Is the interface structure obvious and effective?
2. Are graphical elements (icons, graphs) recognised as expected?
3. Do users understand and/or accept physical therapy assistance?
4. Do users understand and/or accept real time gait feedback?
To investigate the first three aspects, a functional prototype can be used. The functional prototype should at
least allow the user to perform some (interactive) tasks. To investigate the fourth aspect, a functional
prototype of the tactile feedback module is required.

GUI Prototype
A functional prototype of the user interface has been developed 36. The prototype is based on HTML and
Javascript, and allows the user to perform simple tasks. The following functions are supported:
- Simulate real-time feedback - User can start real time feedback mode, and control the rhythm speed
using the PDA. Tactile feedback is simulated graphically.
- Simulate PT feedback - User can start physical training. Two examples have been included. The user
can ask for more information about the exercise, and evaluate the exercise afterwards.
- Show progression - An overview of weekly and monthly progression can be shown. User can select
weekly or monthly view.
- Check system status - The user can check the battery status and connection status of each of the three
system modules
Besides the user actions, the interface prototype also simulates several external conditions:
- Start physical training alarm
- Simulate low battery
- Simulate lost signal
- Simulate incoming message from caregiver
^ |Q fik::)|Jhorne/josJdocBjLJt/:iaar%g0JlfUbiquitaLJS%g0Computinq|5ensor5hoe%20GUI)start,htril |T| ||G|T|’: ■■ : ■ _

Fig. 5: Screenshot of the prototype in a webbrowser

Limitations
The prototype is only meant to be used for communication (showing how the interface would work) and
usability testing purposes. Integration with other prototype parts, such as sensors or tactile feedback

36
Appendix D contains screenshots of the prototype

43
SensorShoe: Mobile Gait Analysis

actuators, isn't feasible with this prototype.

Furthermore, at the moment the prototype only runs on a 'normal' computer. It hasn't been tested on actual
PDA's or other mobile devices. However, it does provide a simulated on-screen PDA. This limitation should
be taken in to account during the usability tests.

Test Setup
To test the usability of the PDA GUI, a group of users can be asked to perform certain tasks with the
interface. An introduction to the subject of the project would be useful to give the test persons an impression
of what they may expect.

Test Group
The test group should represent both general and extreme characteristics of the users, as described in
Chapter 3. For a first test, a group of 15 persons would be a good trade-off between significant quantity and
available time.

Environment & Participants

The usability test can be conducted in a normal environment. The only requirements are a (mobile) computer
with an internet browser and the functional prototype. To receive user input it is possible to use a normal
mouse, but it would be better to use a touch screen if available. This would be a better simulation of the
actual PDA interface.
Besides the test person, a facilitator should be present to provide the user with tasks, and help the user if
needed. The facilitator also introduces the subject and the functional prototype. Furthermore, another
member of the test team should record the actions of the user. This can be done by writing notes, or by
automatically logging the computer input/output actions.
Tasks
Using the functional prototype, the users can be asked to perform the following tasks.
1. Suppose you'd like to make a walk and receive assisting cues. Start the appropriate program on the
PDA.
1. [At thereal time gaitscreen] You're receiving cues. Lower the speed of the cues.
2. [At thereal time gait screen] You've arrived at your destination. Stop the cues.
3. [At thereal time gait screen] Go to the main screen
2. You want to do your daily physical therapy exercises. Start the appropriate program on the PDA.
1. [At the PT screen] Suppose you want more information about the current exercise. Find the
appropriate support on the PDA.
2. [At the PT screen] After finishing the exercise, express your opinion about the exercise
3. You want to check the progression of your exercises. Start the appropriate program on the PDA.
1. [At the progression screen] Which days of the last week were most effective?
2. [At the progression screen]Go back to the main screen

The functional prototype can also be used to simulate external actions.

1. [ Facilitator starts the 'low battery' message ]
1. What's going on?
2. What would you do now?
2. [ Facilitator starts the 'out of range' message ]
3. What's going on?
4. What would you do now?
The tasks and simulated events should be prepared by writing or printing them on cards. The cards can be

44
SensorShoe: Mobile Gait Analysis

used by the facilitator to instruct the test person. The results of the above tasks should provide insight in to
the understanding and effectiveness of the SensorShoe interface.

45
SensorShoe: Mobile Gait Analysis

5.3. Conclusions
This chapter conducted a first evaluation of the SensorShoe concept by reviewing the system requirements
of Chapter 3. The results indicate further development should focus on the following aspects:
1. Communication between the PDA and the caregiver
2. The PDA software
3. The specifications of electronic components
4. The effectiveness of the GUI
The first three aspects are beyond the scope of this research, as they are mostly related to component
design and detailed software design. The fourth aspect, regarding the user interface, is relevant for this
research and was further investigated.
A plan for more comprehensive tests of the PDA GUI has been presented. To conduct this usability test, a
functional prototype should be used. This functional prototype is already available, and allows a user to
perform several interactive tasks with the PDA GUI.
Several recommendations on the test environment, set up and task examples have been presented, which
should result in a useful usability test. The test should at least indicate problems with the understanding and
acceptance of the interface.

46
SensorShoe: Mobile Gait Analysis

6. Literature

6.1. Papers, Publications and Proceedings

1. Fernandez del Olmo, M. et al, "Evaluation of the effect of training using auditory stimulation on
rhythmic movement in Parkinsonian patients—a combined motor and [F]-FDG PET study ",
Department of Medicine-INEF-Galicia, Laboratory of Neuroscience and Motor Control (NEUROcom),
Universidad de A Coruna, Spain , 2005
2. Gutihrrez, E. et al, "Virtual images as a perceptual orthesis to assist Parkinson's disease gait
disorders ", Proceedings of The First Joint BMES/EMBS Conference Senring Humanity, Advancing
Technology, Georgia, USA
3. Han, Iung-Sw et al, "Development of the Gait Assistant Mobile Robot Using Ergonomic Design ",
Department of Precision Mechanical Engineering, Hanyang University, Seoul, South Korea , April
2004
4. Jyrinki , T, "Perception of Small Device Vibration Characteristics - Test Facilities Setup and Study ",
Helsinki University of Technology, October 12, 2004
5. Jones, L. et al, "Development of a Tactile Vest", Department of Electrical Engineering,
Massachusetts Institute of Technology, USA, April 2004
6. Langston, J.W., "PD: More the a Movement Disorder". New York, USA: Parkinson's Disease
Foundation, 2006
7. Lee, J.C et al, "Haptic Pen: Tactile Feedback Stylus for Touch Screens", MITSUBISHI ELECTRIC
RESEARCH LABORATORIES , December 2004
8. Lewis, G.N. et al, "Stride length regulation in Parkinson's disease: the use of extrinsic, visual cues",
University of Auckland, New Zealand , published in "Brain", 123, 2077-2090, 2000
9. Lunenburger et al, Review "Biofeedback for robotic gait rehabilitation ", published in Journal of
NeuroEngineering and Rehabilitation , January 2007
10. Luk, J. et al, "A Role for Haptics in Mobile Interaction: Initial Design Using a Handheld Tactile Display
Prototype ", published in proceedings of ACM Conference on Human Factors in Computing Systems
(CHI'06).
11. Martin JP. Locomotion and the basal ganglia. In: Martin JP. The basal ganglia and posture. London:
Pitman Medical; 1967.
12. Melnick, M.E. et al, "Gait analysis and Parkinson's Disease", published online at "Rehabilitation
management", www.rehabpub.com,August 2002. Last visited May 9 th, 2007
13. Morris, M.E, "Movement Disorders in People With Parkinson Disease: A Model for Physical
Therapy", published in "Physical Therapy", Volume 80, Number 6, June 2000
14. Morris, M.E et al, "The biomechanics and motor control of gait in Parkinson disease", School of
Physiotherapy, La Trobe University, Australia , April 5 2001
15. Morris, S.J, "A Shoe-Integrated Sensor System for Wireless Gait Analysis and Real-Time
Therapeutic Feedback ", Massachusetts Institute of Technology, 2004
16. Paradiso, J.A et al, "Interactive Therapy with Instrumented Footwear ", Responsive Environments
Group, MIT Media Laboratory Cambridge, MA USA 02139 , February 2004
17. Peppe, A. et al, "Does gait analysis quantify motor rehabilitation efficacy in Parkinson's disease
patients?", Fondazione Santa Lucia, Roma, Italy, November 2006
18. Stone, D. et al, "User Interface and Interaction Design", The Open University, Published by Elsevier,
2005

47
SensorShoe: Mobile Gait Analysis

19. Scheibe, R. et al, "Tactile Feedback at the Finger Tips for Improved Direct Interaction in Immersive
Environments , Volkswagen AG, Group Research, January 2007
20. Tsukada, K. et al, "ActiveBelt: Belt-type Wearable Tactile Display for Directional Navigation ",
Graduate School of Media and Governance, Keio University, Japan, June 2004
21. van Wegen, E. et al, "The effect of rhythmic somatosensory cueing on gait in patients with
Parkinson's disease ", published in "Journal of the Neurological Sciences", 248, 2006
22. van Wegen, E. et al, "The effects of visual rhythms and optic flow on stride patterns of patients with
Parkinson's disease ", Department of Physical Therapy, VU University Medical Center (vUmc),
Amsterdam, The Netherlands, June 2005

6.2.Internet Sources
1. "The GaitShoe project", http://www.media.mit.edu/resenv/GaitShoe/index.html, visited in May 2007
2. "Parkinson's Disease at Wikipedia", http://en.Wikipedia.org/wiki/Parkinson's_disease, visited on 22-3-
2007
3. "Parkinson's Disease Foundation", http://www.pdf.org, visited on 22-3-2007
4. "Bluetooth Technology Brief", http://www.silicon-press.com/briefs/brief.bluetooth/index.html, visited
on May 9th, 2007
5. "PDA Specifications Example", http://www.pdashop.nl/product/19068/hp-ipaq-
hx2490.html#specificaties, visited on May 7th 2007
6. "The E-Skin Lab", http://www.e-skin.ch/, visited May 2007
7. "Kinestetic Interfaces based on Hasser" - http://www.hitl.washington.edu/scivw/scivw-
ftp/publications/IDA-ps/HAPTIC2.PS
8. The Frequently-Asked Questions about Parkinson's Disease. New York, USA: Parkinson's Disease
Foundation, 2006, http://www.pdf.org. visited in May 2007

48
SensorShoe: Mobile Gait Analysis

Appendix A - Feedback Selection

Audio

There are three types of audio that can be used for feedback, namely rhythmic earcons, spoken text or
music. The audio can be presented to the user trough a speaker, which may be implemented by the one of
the following techniques.
Speaker Pro's Con's
A speaker can be placed on a
PDA or on the SensorShoe it The user can hear the clues of Other people in the surrounding area of the
self. the SensorShoe directly. user can hear the clues of the SensorShoe:
this can be very annoying.
This solution can be very stigmatising, since
the people around the user will hear
everything. The user can get too
embarrassed to even want to use the
SensorShoe.

Earphone Pro's Con's

To prevent the stigmatising No stigmatising effect, since By wearing an earphone, the user might
effect, the feedback can be only the user can hear the not hear any other sound. In traffic, this
given via a earphone. feedback. can be very dangerous.
No irritation of surrounding
people.

Hearing aid Pro's Con's

This is a device which is
already used for providing No stigmatising effect, since Signals of the SensorShoe and signals from
hearing aid. The device is only the user can hear the the 'outside' may interfere, increasing the
place behind the ear of the feedback. workload of the user.
user and amplifies the No irritation of surrounding
surrounding sounds. people.
The user can still hear
surrounding sounds, since the
hearing is not cut off.

49
SensorShoe: Mobile Gait Analysis

Visual

For visual feedback, two forms of output are used for gait feedback, namely "Visual rhythm cues" and "Visual patterns".
Rhythm cues use the same principle as audio cues. Visual patterns may be in the form of line patterns or 'footstep
patterns'. Several technical implementations are discussed below.

Light signals Pro's Con's

It's a cheap and easy solution
The surrounding people can see the light signals.
The device provides light signals As described above; this can be very annoying and
to the user. This can represent a stigmatising
rhythm to improve the users gait.

Cue Projection Pro's Con's

Visual cues should be easily
The cues can also be projected The surrounding people can see the light signals.
understandable for the user
on the ground. As described above; this can be very annoying and
The user can adjust the gait stigmatising
based on these cues

Virtual Pro's Con's

With a head-up display it is A head-up display is extremely expensive.
possible to create an augmented By adding cues to the users
reality. Virtual elements are reality, its easy to adjust the
added to the reality. gait.
Surrounding user do not see the
augmented reality elements.
The be able to provide a augmented reality, the
user should wear a large system on its head. It is
no possible to make small VR solutions.

A fully working VR solution is not yet feasible.

Computer screen Pro's Con's

When using a portable computer The user always has to look at the PDA while
screen (i.e. PDA) the display can The display is only visible for the walking
be used to provide the cues. user, so there is no stigmatising
effect and it is not annoying for
other users.
The costs (of development and
implementation) are not as high as
a virtual reality solution. This is
feasible already, a VR solution is
not.

A computer screen can provide all

kinds of feedback; icons, colours,
movement on the screen etc.

46
SensorShoe: Mobile Gait Analysis

Tactile

Tactile feedback uses the skin of the user to send information. The information may be a rhythm or a
direction, or a combination of both. Possible technical implementations are presented in the tables below.
Rhythmic taps Pro's Con's

The device on the body of the The technique for this can be A single tap could easily be mistaken for
user provides a rhythmic small and therefore invisible' other movements or touches.
feeling. The user senses the for other people.
taps and adjusts the gait. Easy to adjust gait

Electronic pulses Pro's Con's

Via electrodes an electronic
pulse can be given to provide The technique for this can be Electronic pulses can be very dangerous for
a rhythm. small and therefore invisible' people with weak harts.
for other people.
The technique for this can be Depend on variable factors such as humidity
small and therefore invisible' and skin characteristics.
for other people.

Vibration Pro's Con's

Just like a rhythmic tap, the
vibration is a small feeling After a while the user may be annoyed by
placed on the users skin. Higher chance of being noticed the vibrations.
(compared to rhythmic taps).
The technique for this can be
small and therefore invisible'
for other people.

Movement on skin Pro's Con's

Sliding an object over the The technique for this can be Movements on skin can't be used for
skin can be used to give a small and therefore invisible' setting a certain rhythm.
sense of direction to the for other people.
user

Pinching / Squeezing Pro's Con's

Mechanically pinching or The technique for this can be Pinching could hurt the user.
squeezing the skin small and therefore invisible'
for other people.
Pinching may be used for both After a while the skin may get irritated.
directional and rhythm
cues.

51
SensorShoe: Mobile Gait Analysis

Appendix B - GUI Concept Evaluations

Heuristic Evaluation
Several guidelines, as described in User Interface Design and Evaluation (Stone et all.), will be used for
this37.
General It is often best to follow accepted conventions if your users are
1.1
familiar with these.

Most users read from left to right and from top to bottom. You should 1.2
order widgets to reflect this.
Follow the layout, text, color, and image guidelines explained in Part 1.3
4.

Primary To identify primary windows, start by looking at the main task objects
2.1
windows in the conceptual design
A launch pad window can be a useful way to organize primary
2.2
windows

Secondary Message boxes should be worded so that the user understands the 3.1
windows message. Avoid unnecessary jargon.
Avoid using unnecessary message boxes.. 3.2
3.3
Use modal secondary windows if the situation requires immediate
attention and it is important not to enter any further data.
3.4
Use dialog boxes if additional information is required to carry out a
task.
Menus 4.1
Menu items should be named so that their name indicates their
purpose.
4.2
The menu structure should be organized around the needs of the
users rather than around the underlying software. Card sort can help
achieve this.

Toolbars ToolTips can help the user to understand the meaning of icons 5.1
5.2
Design icons that users can easily recognize and understand.
Design icons that are visually simple. 5.3
Design icons that are informative 5.4
Design icons that can be easily distinguished 5.5
Design icons that represent concrete objects 5.6
Design icons that are easy to perceive 5.7
Stone, D. et al, 2005

52
SensorShoe: Mobile Gait Analysis

Command Commands should be worded so that they clearly indicate the

6.1
buttons action that the button carries out.

Place command buttons along the bottom of dialog boxes or up 6.2

the right-hand side
6.3
The buttons on a dialog box should be the same size and shape.
Different-width buttons are acceptable if the labels are different
lengths and the buttons are in a row

Text boxes Use a text box if it is not possible to anticipate the user input. 7.1
7.2
Do not use a text box without a list box if the GUI requires
standardized information.
7.3
The size of the text box should indicate how much information is
required.
7.4
The text box should be scrollable if it is not possible to anticipate the
quantity of user input.
7.5
If the text box is scrollable, ensure that sufficient lines are visible
to give sufficient context for the person entering the text.
7.6
Gray-out the text box if you want to show that, in a
particular context, the content of the box cannot be
changed.

53
SensorShoe: Mobile Gait Analysis

Principles and guidelines for website design

Because these guidelines were made for a website, only the relevant guidelines for a application on a PDA
will be discussed.
The structure
Take into account the natural organization of the information 8.1
of the site
Make the structure intuitive to your users 8.2

Helping the Create a consistent visual appearance for the site 9.1
users know
where they
are

Site maps

Writing for the

Keep the language simple and avoid jargon 11.1
web
Check your spelling and grammar 11.2

Always include a link to the first screen

10.1

Also these guidelines were made for websites. Therefore, not all of them will be relevant. Only the relevant
guidelines will be discussed here.
12.1
Individual pages of a site should have a consistent and simple layout
so that users with visual impairments or blind users using screen
readers can more quickly navigate through a page and find the
information they are trying to locate.
Backgrounds should be kept simple, with enough contrast so that 12.2
Websites for users users
with disabilities
with low vision, color blindness, or black and white monitors
can read the visual clues.
12.3
Buttons should be large, easy targets so that users with physical and
mobility disabilities can select them easily from the screen.
12.4
Functional features-buttons, scroll bars, navigational bars- should be
identified as working functions rather than images.
12.5
Blinking or constantly changing text elements should not be used, so
that users with visual impairments, learning disabilities, or recurring
headaches are not challenged.

54
 SensorShoe: Mobile Gait Analysis

Evaluations
Each guideline will be code. Firstly, the guideline will be graded with regards to each concept design. The
grades will be between 0 and 2. 0 means "Does not fit the guideline at all" and 2 means "Fits the guideline
fully".
Secondly, the concepts will be judged according to the potential which they have. (if the guideline is not fully
met) The grades they will receive is the difference between the first grade and the potential the concept
design has in the same scale as mentioned above.
The first grades will be colour coded, in order to get a better overview. Green is 2; guideline met. Orange is 1;
guideline partially met. Red is 0; guideline not met.
Since the concept designs are not made into details yet, the grading is based on just the known properties of
the interface. If a concept design lacks functionality, it will not receive the grade in the first guideline grading
part. However, if it has the potential to have this functionality, it will receive the grades in the second guideline
grading part.

Concept 1 Concept 2 Concept 3 Concept 4 Concept 5

2nd 2nd 2nd 2nd 2nd
1 st
1 st
1 st
1 st
1 st

General 1.1 0 0
1.2 0
1.3 0 0
Primary 2.1
windows
2.2 0
Secondary 3.1 - 2 - 0 - 2
windows
3.2 - 2 - 2 - 2
3.3 - 2 - 2 - 2
3.4 - 0 - 2 - 2 - 2 - 2
Menus 4.1 0 0
4.2 0 0 0 0
Toolbars 5.1 1 2 2 0 0
5.2 0 0 0 0
5.3
5.4 0 0 0 0
5.5 0
5.6 0 0 0 0
5.7 0
Command 6.1 0 0 0 0
buttons
6.2 1 1 1 1
6.3

55
 SensorShoe: Mobile Gait Analysis
Text boxes 7.1 - 2 - 2 - 2 - 2 - 2
7.2 - 2 - 2 - 2 - 2 - 2
7.3 - 0 - 0 - 0 - 0 - 0
7.4 - 1 - 1 - 1 - 1 - 1
7.5 - 0 - 0 - 0 - 0 - 0
7.6 - 0 - 0 - 0 - 0 - 0
Total score 23 13 30 9 12 14 23 8 17 14

Explanation when guidelines are not, or partially, met:

Concept 1
3.1,3.2,3.3,3.4: no secondary windows were assigned yet.
5.1: no tooltip is used.
5.2: some of the icons are the same as a tape recorder; the target group might not have great experience
with tape recorders.
5.4: since the icons are simple, they might me too simple.
5.6: the icons do not represent objects but actions.
6.1: commands are only given with buttons, no words.
6.2: command buttons are on the left.
7: text input is not made into details (this is only used in the evaluation function)

Concept 2
1.1: Accepted conventions are not used. This could also be seen as a great advantage. The users will not be
bothered with complex knowledge of other interfaces. This interface is simple and intuitive. The user can start
to use it, without bias.
5.1: ToolTips are not used. These can be added since the interface has space left.
5.2: some of the icons are the same as a tape recorder; the target group might not have great experience
with tape recorders.
5.4: since the icons are simple, they might me too simple.
5.6: the icons do not represent objects but actions.
6.1: commands are only given with buttons, no words.

Concept 3
1.1: the menu structure does not follow accepted conventions at all. There are no known menu's like this. By
changing the structure, it changes the whole menu.
1.2: the header is placed on the right, the text is rotated. Also, the menu items are at the bottom.
1.3: The header text is rotated.
3.1,3.2,3.3,3.4: no secondary windows were assigned yet.
4.1: Menu items at the bottom are not named, there isn't any space to name them either.
4.2: there is no special organization, the menu items are just placed together in order to reduce used space.
5.1: ToolTips are not used. These can be added since the interface has space left.
5.2: The triangle shaped icon is difficult to recognize. It should indicate that the menu can expand, but it
seems to be part of the menu shape.

56
 SensorShoe: Mobile Gait Analysis
5.4: since the icons are simple, they might me too simple.
5.5: the menu icons are all on one bar. They are difficult to distinguish.
5.6: the icons do not represent objects but actions.
5.7: as mentioned at 5.2 and 5.5.
6.1: commands are only given with buttons, no words.
6.2: no dialogue boxes assigned yet.
7: text input is not made into details (this is only used in the evaluation function)

Concept 4
4.2: Menu items are placed in one box to order them. They are not ordered around the needs of the user. 5.1:
no tooltip is used. Difficult to add, there is no space left.
5.2,5.3,5.4,5.5,5.6,5.7: even thou no icons are used, these guideline are not met. The guidelines mean: the
user should understand the menu items. In the toolbar text is used, which is difficult for the target group. Text
is therefore not simple, easy to perceive etc.
6.2: no dialogue boxes assigned yet.

Concept 5
1.3: the font size will be too small. The boxes of the calendar will be are to distinguish since there small as
well.
2.2: the main screen will not exist of a launch pad, but of several calendars.
3.1,3.2,3.3,3.4: no secondary windows were assigned yet.
4.1: the menu uses icons. There is no space for naming tags.
4.2: The remaining functions are located at the bottom of the interface, no special ordering.
5.1: no tooltip is used.
5.2: some of the icons are the same as a tape recorder; the target group might not have great experience
with tape recorders.
5.4: since the icons are simple, they might me too simple.
5.6: the icons do not represent objects but actions.
6.1: commands are only given with buttons, no words.
6.2: no dialog boxes assigned yet.
7: text input is not made into details (this is only used in the evaluation function)
Principles and guidelines for website design (p. 413-415)
Because these guidelines were made for a website, only the relevant guidelines for a application on a PDA
will be discussed.

Concept 1 Concept 2 Concept 3 Concept 4 Concept 5

2nd 2nd 2nd 2nd 2nd
1st
1st
1st
1st
1st

The structure of 8.1 0

the site
8.2 0 0 0 0
Helping the 9.1 0 0 0
users know
where they are

57
 SensorShoe: Mobile Gait Analysis
Site maps 10.1 1 1
0
Writing for 11.1
the web
11.2 - 2 - 2 - 2 - 2 - 2
Total score 6 2 10 2 4 3 7 3 9 2

Concept 1
8.1: the structure is not intuitive, since it looks a lot like standard interfaces, but is different in several
elements. Is can be confusing.
9.1: the background colour are different in each function, this is not consistent.
10.1: a link to the main screen is not always included. This could be included, but the screen will get very
crowded.

Concept 3
8.1: all options are placed on one bar, there is no organization.
8.2: it's difficult for users to find the menu items.
9.1: the same menu structure is not possible for "real time gait support".
10.1: there is no button yet, but it could be added.

Concept 4
8.2: it's not clear that the menu button is clickable.
9.1: since the function will decide the look of a page, the menu is always folded in, there is no consistency.
When the menu is extracted, there is consistency.
10.1: there is no button yet, but it could be added.
Concept 5
8.2: even thou a calendar is intuitive (it's known) many calendar on one page will be confusing. (what to do
with them?)

Websites for users with disabilities (p. 417-418)

Also these guidelines were made for websites. Therefore, not all of them will be relevant. Only the relevant
guidelines will be discussed here.

Concept 1 Concept 2 Concept 3 Concept 4 Concept 5

2nd 2nd 2nd 2nd 2nd
1 st
1 st
1 st
1 st
1 st

12.1 0 0 0 0
12.2
12.3 0 0
12.4 0 0 0 0
12.5 0 0
LD

7 3 7
00

0 0 0 0 0

Concept 1

58
 SensorShoe: Mobile Gait Analysis
12.1: Background colour change.
12.4: the buttons could also be images.
Concept 2
12.4: the buttons could also be images.
Concept 3
12.1: Since the same menu structure is not fully possible for other functions, it's not possible to create a
consistent interface.
12.3: buttons are small.
12.4: the buttons could also be images.
12.5: the menu is moving in the screen.
Concept 4
12.1 See 9.1
12.5: the menu is moving in the screen.
Concept 5
12.1: only the header and buttons will be in the same style. The content style of the page will change a lot in
every screen.
12.3: the buttons of the calendar cannot be large, since the whole calendar has to fit on one screen.
12.4: the buttons could also be images.

59
 SensorShoe: Mobile Gait Analysis
Appendix C - Sensor Specifications

Accelerometer 37

Parameter Conditions Min Typ Max Unit

SENSOR INPUT Each axis
Measurement Range ±3 ±3.6 3
Nonlinearity 9iof full scale ±03 %
Package Alignment Error ±1 Degrees
Interaxis Alignment Error ±0.1 Degrees
Cross Axis Sensitivity’ ±1 %
SENSITIVITY {RATIOMETRIC)1 Each axis
Sensitivity at XOUT* YOUT< ZOUT VS = 3V 270 300 330 m V/g
Sensitivity Change Due to Temperature3 Vs = 3 V ±0.015 %/°C

ZERO g BIAS LEVEL (RATIOMETRIC) Each axis

0 Q Voltage at XOUT,YCXJT< ZOUT Vs = 3 V 1,2 1.5 1.8 V
0 g Offset vs. Temperature ±1 mgFC
NOISE PERFORMANCE
Noise Density Xcur, YOLT 280 pg/i/Hz rms
Noise Density ZDUT 350 pg/i/Hz rms

FREQUENCY RESPONSE4
Bandwidth Xour. YOUTS No external filter 1600 Hz
Bandwidth iurs No external filter 550 Hz

RFIT Tolerance 32 ± 15% kn

Sensor Resonant Frequency S.5 kHz
SELFTEST4
Logic Input Low +0.6 V
Logic Input High +2.4 V

ST Actuation Current +60 pA

Output Change at Xour Self test 0 to 1 -150 mV
Output Change at YOUT Self test 0 to 1 +1 50 mV
O utput Cha nge at Zour Self test 0 to 1 -60 mV

OUTPUT AMPLIFIER
Output Swing Low No load 0.1 V
Output Swing High No load 2.8 V
POWER SUPPLY
Operating Voltage Range 1JB 3.6 V
Supply Current Vs = 3 V 320 UA
Turn-On Time7 No external filter 1 ms
TEMPERATURE
Operating Temperature Range -25 +70 °c
1
Defined as cou pi ing between any two axes.
2
Sensitivity is essentially ratiometric to V*.
’Defined as the output change from ambient-to-maxi mum temperature or ambient-to-minimum temperature.
* Actual frequency response controlled by user-supplied external filter capacitors {C«CVr Cr).
5
Bandwidth with externaI capacitors = 1 /tf x n x 32 kO xC). ForC*Cv = 0.003 pF, bandwidth = 1.6 kHz. For Cx=0.01 pF, bandwidth = 500 Hz. For C*CY,Ca = 10 pF,
Fig. 6: Datasheet for the 3D accelerometer

37
See http://www.analog.com/UploadedFiles/Data_Sheets/ADXL330.pdf
60
 SensorShoe: Mobile Gait Analysis
Gyroscopic Sensors 38

Parameters Conditions Min Typical Max Unit

SENSITIVITY
Full-Scale Range •500 7s
Sensitivity 2.0 mV/7s
Initial Calibration Tolerance -5 +5 %
Over Specified Temperature ±10 %
Nonlinearity Best Fit Straight Line <1 %cf FS
Cross-axis Sensitivity ±2 %
ZERO-RATE OUTPUT
Static Output (Bias) 1.5 V
Initial Calibration Tolerance -300 +300 mV
Over Specified Temperature -300 +300 mV

FREQUENCY RESPONSE
High Frequency Cutoff Internal LPF -90” 140 Hz
LPF Phase Delay 10Hz 4.5 *

MECHANICAL FREQUENCIES
Resonant Frequency X-Axis Gyroscope 10 12 14 kHz
Resonant Frequency Y-Axis Gyroscope 13 15 17 kHz
Frequency Separation X and Y Gyroscopes 3 kHz
NOISE PERFORMANCE
Rate Noise Density 0.014 ■feA'Hz

OUTPUT DRIVE CAPABILITY

Output Voltage Swing Load - 10OkO to V«/2 0.05 Vdd-0 05 V
Capacitive Load Drive 100 pF
Output Impedance 100 □
REFERENCE
Voltage Value 1.23 V
Load Drive 1 mA
Capacitive Load Drive Load directly connected to VREF 100 PF
Power Supply Rejection VDD= 3.0V to 3.3V 1 mV/V
Over Specified Temperature •5 mV
POWER TIMING
Zero-rate Output Settling to+37sec 200 ms
POWER SUPPLY
Operating Voltage Range 3.0 33 V
Quiescent Supply Current 9.5 mA
Over Specified Temperature •2 mA
TEMPERATURE RANGE
Specified Temperature Range 0 to +70 "C
Extended Temperature Range Performance parameters are not applicable -20 to +85 °C
beyond Specified Temperature Range

Fig. 7: Datasheet for 2D gyroscope

38
See http://www.invensense.com/shared/pdf/IDG300_Datasheet.pdf
61
SensorShoe: Mobile Gait Analysis
Appendix D - Interface Prototype
Screenshots of the PDA GUI prototype.

Realtime gait

The main screen Real Time Gait feedback Progression of a particular physical
exercise

Progression over the last week System status Low Battery warning
check

62