Baran: An Interaction-centred User Monitoring Framework

Mohammad Hashemi and John Herbert

Computer Science Department, University College Cork, Cork, Ireland

Keywords:

User Modelling, User Proﬁling, Quality of Experience, User Experience, Human and Computer Interaction.

Abstract:

User Quality of Experience (QoE) is a subjective entity and difﬁcult to measure. One important aspect of

it, User Experience (UX), corresponds to the sensory and emotional state of a user. For a user interacting

through a User Interface (UI), precise information on how they are using the UI can contribute to understanding

their UX, and thereby understanding their QoE. As well as a user’s use of the UI such as clicking, scrolling,

touching, or selecting, other real-time digital information about the user such as from smart phone sensors

(e.g. accelerometer, light level) and physiological sensors (e.g. heart rate, ECG, EEG) could contribute to

understanding UX. Baran is a framework that is designed to capture, record, manage and analyse the User

Digital Imprint (UDI) which, is the data structure containing all user context information. Baran simpliﬁes

the process of collecting experimental information in Human and Computer Interaction (HCI) studies, by

recording comprehensive real-time data for any UI experiment, and making the data available as a standard

UDI data structure. This paper presents an overview of the Baran framework, and provides an example of its

use to record user interaction and perform some basic analysis of the interaction.

1 INTRODUCTION

Interaction is very important and as the number of in-

teractive devices grows, providing a good User Expe-

rience (UX) is a concern for device and service pro-

ducers. Understanding UX is difﬁcult, and it is even

more difﬁcult when UX is for interactive products

and systems (Forlizzi and Battarbee, 2004). There

are several approaches to understanding UX. One is

the perspective of users that is called user-centred,

or focus on products that is called product-centred.

Another is to understand UX through the interaction

between user and products named interaction-centred

(Forlizzi and Battarbee, 2004). Many researchers be-

lieve interaction-centred is the most valuable for un-

derstanding how a user experiences a product (Forl-

izzi and Battarbee, 2004). In this study we focus on

interaction-centred UX.

User Quality of Experience (QoE) is a subjective

entity and difﬁcult to measure. One important aspect

of it, User Experience (UX), corresponds to the sen-

sory and emotional state of a user. For a user interact-

ing through a User Interface (UI), precise information

on how they are using the UI can contribute to un-

derstanding their UX, and thereby understanding their

QoE.

As well as a user’s use of the UI, other real-time

digital information about the user could include the

readings from the various sensors on a device such as

a smart phone (e.g. accelerometer, gyroscope, mag-

netometer, light level) and readings from physiologi-

cal sensors as might be available from a health mon-

itoring wireless BAN (body area network) (e.g. heart

rate, GSR, ECG, EEG, EMG). All this real-time dig-

ital data together is called the User Digital Imprint

(UDI), and it provides all the available external digital

information that can reﬂect the subjective UX when

interacting through a UI.

As the popularity of the mobile devices are in-

creasing, users experience different situations while

working with their mobile devices. User’s context and

moods, and environments are dynamic and always

are changing. Applications, services, and all com-

puting trends need to know the changes and adapt to

them in order to improve user experience. Context is

any information characterizing human and computer

interaction and the surrounding environment. This

work addresses a gap between applications or services

and the context. Each service or application needs

to implement a component to extract some informa-

tion in order to know the context. First of all, this is

a time consuming process and needs special knowl-

edge about context information. Secondly, process-

ing, analysing, and extracting some useful informa-

Hashemi M. and Herbert J..

Baran: An Interaction-centred User Monitoring Framework.

DOI: 10.5220/0005239600520060

In Proceedings of the 2nd International Conference on Physiological Computing Systems (PhyCS-2015), pages 52-60

ISBN: 978-989-758-085-7

 2015 SCITEPRESS (Science and Technology Publications, Lda.)

tion out of the raw data about the context is another

difﬁcult process. Service producers and application

developers should not also have to implement the con-

text component part. Baran is an interaction-centred,

user monitoring framework that meets this need. It

collects context information such as interaction, sen-

sor, and physiological data. In future work, the col-

lected raw data will be analysed and higher level in-

formation will be provided. It currently supports the

Windows Operating System (OS) for desktops and the

Android OS for smart phones and tablets. Mac OS

and iOS versions are under development. The Win-

dows OS version of the framework was presented and

published in (Hashemi and Herbert, 2014).

2 BACKGROUND

There are other systems to collect and manage context

information for mobile devices include middle-ware,

services, and frameworks. CONTextfactORY (Con-

tory), is a middle ware prototype proposed by Riva

(Riva, 2006). It collects some internal and external

context information and offers a SQL-like interface

using a query language to allow third parties to make

a query about context with some speciﬁcations, spe-

cially in an ad-hoc network manner. This work suf-

fers from resource constrained of smart phones and

tablets. Storing the context data needs huge storage

space and also the CPU, RAM, and battery usage for

processing. We address this part by sending the con-

text data to our cloud server in order to store and pro-

cess. We use a well-designed data structure, UDI, for

the data sent to our cloud server. It is explained in

section 4.1.

Hermes, is another context-aware application de-

velopment framework (Buthpitiya et al., 2012). It has

local and cloud service in order to provide context. It

deﬁnes widgets that are responsible for sensors. Wid-

gets can communicate to each other in order to trans-

mit context data. While they address the weaknesses

of some other work, they did support other develop-

ers to use their framework. They also did not provide

samples of their work. Their framework also lacks the

functionality of connecting to BAN sensors.

Some other context systems exist, such as the

Context Toolkit (Dey et al., 2001) that is a library,

RCSM (Yau and Karim, 2004) that is middle ware,

and the TEA framework (Schmidt et al., 1999). They

provide application developers with uniform context

abstractions but mostly without the analysing or pro-

cessing data.

In this study we propose the Baran framework.

We implement it in Java for Android and in C#.Net

for Windows OS. Baran is able to collect the context

information and combine it with BAN sensor data in

order to provide the opportunity to estimate UX. In

future we will work on processing and analysing data

to provide the higher level context information.

The term, User Experience (UX), was introduced

in the 1990s (Forlizzi and Battarbee, 2004) and refers

to a user’s perceived experience of a service or a prod-

uct. UX includes all aspects of behaviour, emotions,

and attitudes. Cognitive, emotional, aesthetic, physi-

cal, and sensual experiences contribute to a user’s ex-

perience. UX is deﬁned as : a person’s perceptions

and responses that result from the use and/or antici-

pated use of a product, system or service (ISO 9241-

110:2010) (9241-210:2010., 2010). UX is subjective,

dynamic, difﬁcult to measure, and also depends on an

individual’s perceived experience and context. The

context could be either external, about the environ-

ment of a user, or his/her internal states, such as mo-

tivation, needs, or mood.

There are varies methods for UX evaluation and

measurement. Questionnaires, interviews, and sur-

veys are used in HCI studies (Vermeeren et al., 2010).

A questionnaire contains a set of questions about

the product, its usability, UX metrics, and the user’s

internal states. One of the most popular and reliable

proposed methods in this area is NASA-TLX (Cao

et al., 2009). Mental, physical, and temporal demand,

performance, effort, and frustration level are all mea-

sured in this method. It relies on the workload of a

task and the other measurements mentioned before. It

has been used in variety of studies (Hart, 2006). How-

ever, there is a need for entry of information by users.

AttrakDiff is another popular questionnaire de-

signed to measure hedonic stimulation, identity and

pragmatic qualities of a product (Hassenzahl, 2005;

Hassenzahl and Tractinsky, 2006). AttrakDiff con-

tains 28 questions. The AttrakDiff questionnaire was

used in a study of exploring the effects of hedonic and

pragmatic aspects of goodness and beauty of differ-

ent MP3 player skins (Hassenzahl, 2008). AttrakDiff

is also used in another study (Schrepp et al., 2006)

on the inﬂuence of hedonic quality on attractiveness.

There are three different interfaces evaluated by 90

people, who received an invitation email in order to

participate.

These methods require users to ﬁll up question-

naires, attend to interview sessions, etc. Complicated,

difﬁcult, and confusing questions in an interview or a

questionnaire can make it unpleasant for users. It is

also not a good user’s internal state indicator as deter-

mining emotions and moods are difﬁcult.

In Human and Computer Interaction (HCI) stud-

ies, understanding a user is the main challenge. It

Baran:AnInteraction-centredUserMonitoringFramework

Figure 1: Baran Framework Overview.

requires knowing what has been done by a user be-

fore and after an interaction. So, there is a need to

provide a more complete set of information than that

from a set of general questions. On the other hand,

the methods used to collect the information need to

be unobtrusive and transparent to the user.

The Baran framework is designed to address the

mentioned challenges and improve the UX measure-

ment methods by providing a more complete version

of the information about a user. It also is building

a general solution to collect, analyse, and process

context information, in order to be used by other re-

searchers, developers, or service producers.

3 BARAN FRAMEWORK

The proposed framework, Baran, transparently, and

unobtrusively collects environmental, behavioural,

and physio-psychological context information about a

user. Figure 1 is an overview of the Baran framework

showing the connectivity of internal and external sen-

sors with its engine. The framework can be used

as a library in a software product, or installed sepa-

rately on a computer or a smart phone. The frame-

work provides the ability of connecting to sensors

such as EEG, GSR, EMG, or ECG. Thanks to com-

mercial versions of these sensors, there are plenty of

them available. They are small in size and can trans-

fer data over Wiﬁ and Bluetooth. When a sensor is

conﬁgured, the framework synchronizes the data from

it with the context information, explained in the next

section, and stores them in a well-designed data struc-

ture called the User Digital Imprint (UDI). A collec-

tion of UDI information makes a proﬁle for a single

user. In Baran, a cloud service receives all UDI from

the experimental devices. The Baran cloud service

stores all UDIs, and later provides them for analy-

sis. The analysis methods can be chosen by the re-

searcher. Baran will also provide analysis methods

in the future. As the UDI contains low level infor-

mation, there is a need to process it to extract some

higher level information such as user emotional or

movement state. Having all information in one place

Figure 2: Baran Framework High Level Architecture.

will provide an opportunity to share experimental in-

formation anonymously with other researcher. As the

data is in the cloud, third parties do not need to have

any connection to the user devices in order to receive

context information. They can communicate directly

with the Baran cloud server and after analysing the

context, can accordingly change their applications or

services.

Figure 1 shows interactions through devices, com-

puters, smart phones, or tablets, with services or prod-

ucts, specially through user interfaces (UI). Baran

gathers available data and put them together to build

a UDI. The UDIs are sent to the Baran cloud service

or stored locally.

4 BARAN ARCHITECTURE

The Baran framework contains two main components

as it is shown in Figure 2. One is a service work-

ing transparently as a background process on the de-

vice regardless of being a smart phone, a tablet, or

a computer machine. It collects context information

and builds the UDI, then periodically sends it to the

Baran cloud service. The data structure needs to be

light enough due to have a limited network resources.

It dynamically changes in size, depends on what de-

vice it was built on and what information was avail-

able at the time of collecting data. It encourages us to

use Java Script Object Notation (JSON) format, be-

cause it is a well-designed data structure for dynamic

data. Baran framework builds a multi purpose data set

of this useful information. Baran provides a general

solution giving an opportunity to compare a result of

a study to another.

4.1 User Digital Imprint (UDI)

A User Digital Imprint (UDI) is the digital imprint of

a user that contains a combination of the context data,

UI activities, and the sensor data. UDI, is a data struc-

ture used by the Baran framework, can contains UID

PhyCS2015-2ndInternationalConferenceonPhysiologicalComputingSystems

Figure 3: UDI Data Structure.

UID: User Interface Interaction Data

CD: Internal Context Data

ECD: External Context Data

Figure 4: Baran UDI Collector Engine.

(User Interface Interaction Data), ICD (Internal Con-

text Data), and ECD (External Context Data). UDI

varies in size depending on the device and the con-

ﬁguration of the framework used in an experiment.

Along with UID, ICD, and ECD data, the information

about the device such as the OS, the Model, and the

ID will be added to the UDI. Figure 3 is an overview

of a sample UDI.

4.2 Baran UDI Collector

Figure 4 is the internal architecture of the Baran UDI

collector. It works as a background process, transpar-

ent to the user. It builds a UDI per second as well as

building a UDI for every single UI interaction such

as touching, scrolling, and sliding in the touch screen

devices, and clicking, and typing in the other devices.

The Baran UDI Collector contains four modules, UI

activity, the Internal, the external Context modules,

and a module to build the synchronized version of

the UDI. The UDI contains different information de-

pends on the device and the framework conﬁguration.

For example, desktop computers normally have no ac-

Figure 5: User Interaction Data (UID).

Figure 6: Internal Context Detector.

celerometer or gyroscope which could be added to the

Baran Framework as external sensors. The UI activ-

ity detector module is responsible for detecting the UI

interactions such as clicking and touching. It extracts

important information about the detected interaction

such as the name of application, the name and the type

of the user interface element that was interacted, also

the event identiﬁers such as clicking, scrolling, touch-

ing, double-touching. Figure 5 shows an example of

a part of the UDI describing the UI interaction data

(UID).

4.2.1 Internal Context Detector

The Internal Context Detector (Figure 6) is a mod-

ule of the Baran UDI collector. It is responsible to

manage the internal sensors of the device. Normally,

it is useful for smart devices such as smart phones

and tablets, as they have internal sensors such as ac-

celerometer, gyroscope. Figure 6 lists some available

sensors in some of the smart devices. This module

extracts the raw data from the mentioned sensors and

builds a data structure, called the Internal Context

Data (ICD), to be passed to the UDI Builder. The

Baran framework provides a set of processing and

analysis algorithms to extract high level information

out of the low level information. A user movement

state (eg. Walking, Running, Still) or emotional state

(eg. Happy, Disgusted, Tired) are examples of high

level information (Hashemi and Herbert, 2014).

4.2.2 External Context Detector

The External Context Detector is a module similar

to the Internal Context Detector. Figure 7 shows the

components and the sensors this module is connected

to. This part will able the Baran framework to connect

to external sensors that are not available in the smart

devices. They could be sensors to monitor a user’s

physiological and psychological states. Combining

this data and the information from other components

of the Baran UDI Collector provides us the opportu-

nity of better understanding users and their activities.

Baran:AnInteraction-centredUserMonitoringFramework

Figure 7: External Context Detector.

Some of the important sensors that our framework are

capable with are listed and explained.

• Galvanic Skin Response (GSR): It is also rec-

ognized as Electro Dermal Response (EDR),

Psycho-Galvanic Reﬂex (PGR), Skin Conduc-

tance Response (SCR) or Skin Conductance Level

(SCL), and is as a measurement of the electrical

conductance of the skin and shows the level of the

stress and the emotional state of a user.

• ElectroCardioGraphy (ECG): It measures electri-

cal impulses from the heart and converts them to

a wave form.

• ElectroEncephaloGraphy (EEG): It is a measure-

ment method of brain electrical activity.

• ElectroMyoGraphy (EMG): It detects the electri-

cal potential generated by muscles.

• Facial Expression It reﬂects the emotional state

of a user. Capturing images of the face of a user

helps to detect their emotional state and facial ex-

pression.

This module builds a data structure that is called

the External Context Data (ECD) after collecting all

required data from the external sensors and devices.

The ECD will be passed to the UDI Builder module.

5 DATA COLLECTION AND

ANALYSIS

We use the Baran framework to collect some UDI data

to demonstrate some basic analysis and data variation.

After the data collection process, it was transferred

to the database of the framework, analysed, and pro-

cessed for the demonstration. Five user devices are

chosen for this experiment. Baran was installed on

the device as a service and users had worked normally

with their devices with no distraction. Baran captured

the data, built UDIs, and stored them locally on the

Table 1: Devices Information and Experiment Duration.

Device

Android OS Version

Brand

Model

Type

Dev1 4.4.2 Asus Nexus 7 Tablet

Dev2 4.3.0 Samsung GT-I9300 S-Phone

Dev3 2.3.5 HTC Explorer S-Phone

Dev4 4.1.2 Samsung GT-I8190N S-Phone

Dev5 4.2.2 Sony CS2305 S-Phone

Table 2: User Interaction Usage.

Devices Day : Hour : Minute : Second

Dev1 01:03:04:29

Dev2 01:02:05:15

Dev3 00:00:06:19

Dev4 00:02:08:51

Dev5 02:01:13:30

memory of the device as well as sending UDIs to the

Baran Cloud Service. It is vital to mention that the ex-

periment is to show a sample of interaction data and

experiments with more sensors, EEG, GSR, etc., are

under way.

Table 3: User Interface Interactions (Every Second Data not

included in Total Per Device.

Devices

Touch/Click

Scroll

View Changed

View Selected

Every Second

Total Per Device

Average Interaction Per Day

Dev1 437 890 18 16 76355 1389 138

Dev2 197 574 30 50 - 893 223

Dev3 45 - 3 18 - 404 202

Dev4 194 170 413 16 8249 850 170

Dev5 667 7367 1957 124 120337 10452 1045

Table 1 provides information about the devices

used in this experiment. They are the Android smart

phones or tablets. Users are researchers, age ranged

30 to 35 years old. Table 2 shows the total time users

interact with the devices and Baran collects UDIs. Ta-

ble 3 shows a summary of users interaction such as

touching and scrolling that are the basic interactive

gestures. Other events are deﬁned as View Changed,

that is the event when the user interface elements

changed, View Focused, that is trigged when a user

interface is activated and comes to the front, and Ev-

PhyCS2015-2ndInternationalConferenceonPhysiologicalComputingSystems

100

150

200

dev1 dev2 dev3 dev4 dev5

Devices

Number Of Touches

Apps

chrome

Facebook

Launcher

settings

Viber

Figure 8: Touches/Clicks Summary for the Top Three Ap-

plications Per User.

ery Second, that is an event designed by the Baran

framework to capture the sensor data every second in

order to have a better overview of the contextual infor-

mation. User ﬁve’s data shows that he/she uses his/her

smart phone more than others. Although the duration

of the data collection process for him/her was more

than others, however the average interaction per day

metric also shows a considerable difference.

Figure 8 reports the number of interaction in terms

of touching or clicking the user interfaces of different

applications. Launcher is the name of the operating

system UI manager also called ”Home”. It is obvi-

ous the it is used mainly by users one to four. Device

setting also is used many times bye some users that

shows it is an important part od a smart phone. Face-

book is used by user one and ﬁve as it is in top three

most touched applications. User ﬁve mainly used

Viber, Facebook and Chrome. Chrome, which is a

web browser application, is also used by user two and

ﬁve as the second most touchable application. Viber

is an internet messaging and calling application, that

is used bye some of the users. It is used by user three

and ﬁve in most touchable applications. It shows this

application is used mainly for messaging because the

touching occurs when a user open an application or

typing.

Figure 9 shows the number of scrolling for the top

three applications used by users. The device three

does not support scrolling events and is not reported

in this ﬁgure. This ﬁgure shows that user ﬁve mainly

used chrome with scrolling gesture. Facebook is the

second most scrollable application that user ﬁve used.

Chrome and Facebook are also used by other users as

the top scrollable applications which shows scrolling

within the user interfaces of these two applications

is important and needs to be considered as a key

point in UI designing. Comparing Figure 8 and 9

would help designers to distinguish between toucha-

2000

4000

dev1 dev2 dev4 dev5

Devices

Number Of Scrolls

Apps

chrome

contacts

Facebook

K9Mail

launcher

settings

Viber

Figure 9: Scrolling Summary for the Top Three Applica-

tions Per User.

5000

10000

15000

20000

25000

dev1 dev2 dev3 dev4 dev5

Devices

Duration

Apps

browser

Calendar

chrome

contacts

Disney

Facebook

FRD

GSF

K9Mail

maps

phone

PVZ2RowGame

TTS

Viber

Figure 10: Time Usage Summary for the Top Four Appli-

cations Per User.

bility and scrollability of their applications in order to

improve UX and UI functionality.

Figure 10 shows how long users spend on using

the applications. It reports top four applications for

each user. There is a game played by user ﬁve that is

the main application he/she used. Moreover, there are

some applications that are used in common between

users such as Chrome, Facebook, Contacts, Map ap-

plication.

Figure 11 shows the difference between number

of touching and scrolling events for top ten applica-

tions used by users in total. It is clear that scrolling

is the most popular gesture in the Chrome, the Face-

book, and also some other applications. On the other

hand, touching is mostly used in the Maps and the

whatsApp application. We also compare these top ten

applications in terms of time duration users spend in

an application.

Figure 12 reports that Chrome, Contacts, and

Home are the top three applications that are used by

users in total. There is an interesting observation that

although Facebook application is not used for a con-

siderable amount of time, however the number of in-

Baran:AnInteraction-centredUserMonitoringFramework

2000

4000

6000

Baran

Camera

chrome

Facebook

Home

Launcher

maps

Music

settings

Viber

youtube

Apps

Quantity

Events

Scrolling

Touching

Figure 11: UI Interactions Summary for the Top Ten Appli-

cations for All User.

0e+00

1e+05

2e+05

chrome

contacts

Facebook

Home

K9Mail

maps

PVZ2RowGame

Viber

youtube

Apps

Duration

Apps

chrome

contacts

Facebook

Home

K9Mail

maps

PVZ2RowGame

Viber

youtube

Figure 12: Time Usage Summary for the Top Ten Applica-

tions for All User.

teractions in it, is always high. This data can be used

in order to model users and track their activities to

study abnormal behaviour.

In the smart-phone and tablet OS, each entity in-

teracted with has a title and possibly a content as a

description. Table 4 lists the titles of these entities

that are interacted with during the experiment. For in-

stance, user one in the device one, touches an entity

titled ”Home” more than 100 times. The entity could

be an application named ”Home” or any touch that

was done in the home screen of the device. Analysing

this data needs a more sophisticated methods to ex-

tract useful information. Table 5 reports information

from the accelerometer, gyroscope, orientation, and

rotation sensors. It shows the average and standard

deviation of the sensor readings per device. They

need to be processed in a more sophisticated way to

extract better and useful information. For example,

they could be analysed to ﬁnd the applications that

a user was using on the move. It can be understood

Table 4: The Titles of the Entities Interacted by Users.

Device 1 Device 2

Content

Interactions

Content

Interactions

[Home] 109 [TouchWiz Home] 47

[Apps] 92 [Chrome] 27

[Subway Surf Game] 67 [K9 Email] 22

[Gmail] 33 [Apps] 16

Device 3 Device 4

Content

Interactions

Content

Interactions

[HTC Sense] 35 [TouchWiz Home] 31

[Phone] 16 [unlock,no missed events.] 20

[All apps] 7 [WhatsApp] 20

[End Call] 5 [Apps] 8

[Calendar] 3 [Flipps] 3

Device 5

Content

Interactions

[Facebook] 69

[PvZ 2] 53

[Maps] 49

[Chrome] 47

[Viber] 43

from the accelerometer sensor data that the devices

two and three, were used more on the move.

6 CONCLUSION

Baran is a framework that simpliﬁes the process of

the data collection during an experiment. It is very

valuable to have the contextual information combined

with the User Interaction data. It provides a better un-

derstanding of user behaviour during an experiment.

It also provides a better understanding of the coop-

eration between internal UI elements and can ﬁnd

where a bug occurs. The proposed framework gives

the ability of collecting useful information. This pa-

per demonstrates results of experiments that show the

complexity of the collected data and the variety of the

results obtained from different users. There is a huge

space to add sophisticated analysis algorithms to the

framework to extract more information. It is also very

valuable to be able to combine the contextual infor-

mation with various sensors, such as EEG or ECG.

The Baran framework can be used by researchers

without implementation effort. Researchers can sim-

ply conﬁgure the framework, then the needed data

PhyCS2015-2ndInternationalConferenceonPhysiologicalComputingSystems

Table 5: Average and Standard Deviation of Some Sensors.

Axis X Y Z

DeviceProﬁleID

Average

Standard Deviation

Average

Standard Deviation

Average

Standard Deviation

Accelerometer Sensor Data

Dev1 -4.4900 0.5313 0.8105 2.1964 9.6616 1.1105

Dev2 -3.8800 1.3853 5.8777 2.5352 7.2268 1.9845

Dev3 -0.5919 2.0480 4.8383 3.226 6.7858 3.5863

Dev4 -0.2411 1.0037 2.5825 2.8138 9.1920 1.3464

Dev5 5.8751 3.8237 -1.1253 3.1923 -0.8628 5.0985

Gyroscope Sensor Data

Dev1 -2.50 6.3103 2.80 8.2105 6.099 8.1606

Dev2 4.403 0.3170 -2.05 0.2812 1.32 0.2384

Dev3 0 0 0 0 0 0

Dev4 -9.104 0.184 -1.01 0.1126 -4.509 0.2666

Dev5 0 0 0 0 0 0

Orientation Sensor Data

Dev1 178 130 -5 14 0 3

Dev2 195 144 -37 18 -2 8

Dev3 0 0 0 0 0 0

Dev4 201 87 -15 22 -1 7

Dev5 211 34 89 94 -40 24

Rotation Sensor Data

Dev1 2.58 7.8603 -2.07 9.8902 -0.1908 0.4541

Dev2 0.2871 0.1388 -1.2 0.1617 -4.403 0.332

Dev3 0 0 0 0 0 0

Dev4 9.1704 0.1434 2.820 0.1361 0.2388 0.7033

Dev5 0 0 0 0 0 0

will be collected and provided to them for any further

analysis. Health care system, habit discovery, or ab-

normal behaviour studies are some application areas

the Baran framework and the UDI can be successfully

used. Users also can use it to have a monitoring sys-

tem and improve their UX by ﬁnding good and bad

habits. Moreover, the data is valuable for researchers

who study context based systems (Moldovan et al.,

2013; Sourina et al., 2011). This framework also able

researchers to share data for analysis in science com-

munity. UDI meets all this needs in order to con-

tribute to the science community and help researchers

and developers.

REFERENCES

9241-210:2010., I. D. (2010). Ergonomics of human system

interaction - part 210. Technical report, International

Standardization Organization (ISO). Switzerland.

Buthpitiya, S., Luqman, F., Griss, M., Xing, B., and Dey,

A. K. (2012). Hermes – A Context-Aware Application

Development Framework and Toolkit for the Mobile

Environment. In Advanced Information Networking

and Applications.

Cao, A., Chintamani, K., Pandya, A., and Ellis, R. (2009).

Nasa tlx: Software for assessing subjective mental

workload. Behavior Research Methods, 41(1):113–

117.

Dey, A., Abowd, G., and Salber, D. (2001). A Conceptual

Framework and a Toolkit for Supporting the Rapid

Prototyping of Context-Aware Applications. Human-

Computer Interaction, 16(2):97–166.

Forlizzi, J. and Battarbee, K. (2004). Understanding experi-

ence in interactive systems. In Proceedings of the 5th

Conference on Designing Interactive Systems: Pro-

cesses, Practices, Methods, and Techniques, DIS ’04,

pages 261–268, New York, NY, USA. ACM.

Hart, S. G. (2006). Nasa-task load index (nasa-tlx); 20

Baran:AnInteraction-centredUserMonitoringFramework

years later. Proceedings of the Human Factors and

Ergonomics Society Annual Meeting, 50(9):904–908.

Hashemi, M. and Herbert, J. (2014). Uixsim: A user inter-

face experience analysis framework. In 2014 Fifth In-

ternational Conference on Intelligent Systems, Mod-

eling and Simulation, 2014 Fifth International Con-

ference on Intelligent Systems, Modelling, and Simu-

lation, pages 29–34. IEEE.

Hassenzahl, M. (2005). The thing and i: Understanding

the relationship between user and product. In Blythe,

M., Overbeeke, K., Monk, A., and Wright, P., editors,

Funology, volume 3 of Human-Computer Interaction

Series, pages 31–42. Springer Netherlands.

Hassenzahl, M. (2008). The interplay of beauty, goodness,

and usability in interactive products. Hum.-Comput.

Interact., 19(4):319–349.

Hassenzahl, M. and Tractinsky, N. (2006). User experience

a research agenda. In March-April 2006, pages 91–97.

Moldovan, A.-N., Ghergulescu, I., Weibelzahl, S., and

Muntean, C. (2013). User-centered eeg-based multi-

media quality assessment. In Broadband Multimedia

Systems and Broadcasting (BMSB), 2013 IEEE Inter-

national Symposium on, pages 1–8.

Riva, O. (2006). Contory: A middleware for the provi-

sioning of context information on smart phones. In

Proceedings of the ACM/IFIP/USENIX 2006 Interna-

tional Conference on Middleware, Middleware ’06,

pages 219–239, New York, NY, USA. Springer-Verlag

New York, Inc.

Schmidt, A., Aidoo, K. A., Takaluoma, A., Tuomela, U.,

Laerhoven, K. V., and de Velde, W. V. (1999). Ad-

vanced Interaction in Context. HUC ’99 Proceedings

of the 1st international symposium on Handheld and

Ubiquitous Computing, pages 89–101.

Schrepp, M., Held, T., and Laugwitz, B. (2006). The inﬂu-

ence of hedonic quality on the attractiveness of user

interfaces of business management software. Interact.

Comput., 18(5):1055–1069.

Sourina, O., Liu, Y., Wang, Q., and Nguyen, M. K. (2011).

Eeg-based personalized digital experience. In Pro-

ceedings of the 6th International Conference on Uni-

versal Access in Human-computer Interaction: Users

Diversity - Volume Part II, UAHCI’11, pages 591–

599, Berlin, Heidelberg. Springer-Verlag.

Vermeeren, A. P. O. S., Law, E. L.-C., Roto, V., Obrist,

M., Hoonhout, J., and V

anen-Vainio-Mattila, K.

(2010). User experience evaluation methods: Cur-

rent state and development needs. In Proceedings of

the 6th Nordic Conference on Human-Computer Inter-

action: Extending Boundaries, NordiCHI ’10, pages

521–530, New York, NY, USA. ACM.

Yau, S. S. and Karim, F. (2004). A context-sensitive middle-

ware for dynamic integration of mobile devices with

network infrastructures. Journal of Parallel and Dis-

tributed Computing, 64(2):301–317.

PhyCS2015-2ndInternationalConferenceonPhysiologicalComputingSystems