COMPLEX MOBILE USER ADAPTIVE SYSTEM FRAMEWORK FOR

MOBILE WIRELESS DEVICES

Ondrej Krejcar
Department of Measurement and Control, Center for Applied Cybernetics,
Faculty of Electrical Engineering and Computer Science
VSB Technical University of Ostrava
17. Listopadu 15, 70833,
Ostrava-Poruba, Czech Republic
[email protected]

ABSTRACT
Paper describes a concept of User Adaptive System (UAS) as well as Predictive
Data Push Technology (PDPT) Framework and Biotelemetrical Monitoring
System (BMS) as two joined parts of complex UAS framework. Main focus is in
contribution of UAS to user or patient and his life quality. A Position Oriented
Database on a server and mobile devices is described as important part of whole
UAS, because the position and context of user are one of the most important areas
of UAS. Also the problem of low data throughput on mobile devices is described,
which can be solved by PDPT framework. Localization and user tracking is
described only as a necessary condition for prebuffering realization because the
PDPT Core makes a decision when and which artifact need to be prebuffered. The
accessing of prebuffered data artifacts on mobile device improve the download
speed and response time needed to view large multimedia data.

Keywords: User Adaptive System, Localization, Biotelemetry, Position Oriented

Database, Prebuffering.

1 INTRODUCTION can imagine to have in his mobile UAS is there. A

one specific kind of problems is based in increased
The idea of User Adaptive Systems (UAS) data amount in new mobile systems. In current
grown from interaction between user and system cases, the user need to specify a data to be
(e.g. throws his mobile device). Such interaction downloaded to his mobile device and he need to
can behold in the reaction on user's non declared wait for data downloading and displaying. Due to a
requests. These requests are based on current user several limitations in hardware of current mobile
environment and biological or emotional state (e.g devices, the use of such large amount data has
where I am?, what I feel?, am I ok?, etc.). Such user result in lower user comfort. The needs of any
questions can be answered by sensors on user body techniques to reduce such large data amount or to
or inside the user devices. By the help of user preload them before user’s needs, is still growing
mobile device, we can get a user location (e.g. user up. We created a Predictive Data Push Technology
current position, user future-predicted position, his (PDPT) Framework to solve these problems by data
movement and tracking, etc.). Biomedical sensors prebuffering. Our idea can be applied on a variety
on user body can detect several important of current and future wireless network systems.
biomedical data, which can be used for More usability of PDPT grows from definition of
determination of user emotional state in the area to be prebuffered as well as from evaluation of
environment around. By the combination of user’s artifacts or other user's behaviour sources.
requests (known or predicted) in conjunction with Additional will be presented in sections (3), (6).
other sources of user’s knowledge and behaviours, The second area of problems which we would
the sophisticated information system can be like to solve is based on a users biomedical data
developed based on presented UAS Framework. inputs and a wide area of their possible utility.
The impact of UAS can be seen in the Current body sensors allow a monitoring of a huge
increased user comfort when accessing these number of biomedical data information (e.g. use a
mobile UAS. In ideal case, everything what user special t-shirt equipped with an ECG, temperature,
pressure or pulse sensors). Current hi-tech mobile main types of smart design also overlap between.
devices are equipped with a large scale display, Smart device can also support some type of smart
provide a large memory capabilities and a wide interaction. Smart mobile device can be used for
spectrum of network standards plus embedded GPS control of static embedded environment devices.
module (e.g. HTC Touch HD, HD2). These devices Smart device can be used to support the virtual
have built-in also a special accelerometer which can view points of smart personal spaces (physical
be used to determine a user’s body situation (user is environment) in a personal space which
staying or lying). Last but not least equipment is a surrounding the user anywhere.
light sensor which can be used not only to Satyanarayanan [3] has presented different
brightness regulation. Use of these declared inputs architectures for developing UbiCom systems in
will be discussed and presented in section (4). way of which angle it is focused on a design:
1. Mobile distributed systems are evolved from
2 ARCHITECTURAL DESIGN FOR distributed systems into ubiquitous computing
UBIQUITOUS COMPUTING SYSTEMS 2. UbiCom systems are developed from smart
spaces characterized by invisibility, localized
Ubiquitous Computing (UbiCom) is used to scalability and uneven conditioning.
describe ICT (Information and Communication Poslad [1] has extended a Satyanarayanan
Technology) systems that enable information and model to Smart DEI model (Device Environment
tasks to be made available everywhere, and to and Interactions). Poslads model also incorporates
support intuitive human usage, appearing invisible smart interaction. Smart DEI model also reverses to
to the user [1]. hybrid models. It is widely assuming by users that
Three basic architectural design models for the general purpose of end-user equipment will
UbiCom system can be divided to smart devices, endure but also it will evolve into a more modular
smart environment and smart interaction. The form.
concept of “smart” means that the object is active,
digital, networked, can operate autonomously, is 2.1 Smart Devices
reconfigurable and has a local control of the
resources which it needs such as energy, data A smart device is a device that is digital, active,
storage, etc [1]. These three main types of system networked, user reconfigurable and that can operate
design may also contain sub-systems, sub-parts or to some extent autonomously. Smart devices can be
components at a lower level of granularity that may characterized like personal computers or mobile
also be considered as a smart (e.g., a smart phones with tend to be multi-purpose ICT devices.
environment device may contain smart sensors and These devices operate as single portals used to
a smart controller, etc). An example of a three main access a multiple application services which are
types of UbiCom models is presented in (Fig. 1). running locally on the device or remotely on servers.
A range of forms are available for smart devices.
Smart devices can be defined as personal devices,
having a specified owner or user. In the smart
device model, the place of application user interface
is on side of the smart device. The main
characteristics of smart devices consist of concept
of: mobility, dynamic service discovery and
intermittent resource access.

Figure 1: Three models of ubiquitous computing:

smart devices, smart environments and smart
interaction [1].

Many sub-types of smarts for each of the three Figure 2: Three models of ubiquitous computing:
main types of smarts can be recognized. These smart devices, smart environments and smart
interaction [1]. interact to reach common goals. Components
interact to reach goals jointly because they are
2.2 Smart Environments deliberately not designed to execute and complete
sets of tasks to reach goals all by themselves. There
A first definition of a smart environment are several benefits to designs based upon sets of
brought by Coen [4] as a computation which is interacting components. Interaction between
easily used to enhance ordinary activities. Cook and UbiCom system components does not exist only in
Das [5] refer to a smart environment as ‘one that is one predefined level but it is spread in a range of
able to acquire and apply knowledge about the levels from primitive to smart. Primitive interaction
environment and its inhabitants in order to improve uses fixed interaction protocols between two
their experience in that environment’. A smart statically linked dependent objects. While the smart
environment consists of a set of networked devices interaction uses richer interaction protocols
that have some connection to the physical world. between multiple dynamic independent objects.
The devices which are used for a smart
environment usually execute a single predefined
task (e.g., motion or body heat sensors coupled to a
door release and lock control). Embedded
environment components can be designed to
automatically respond to interaction with user using
iHCI (implicit Human Computer Interaction). A
person can for example walk towards closed doors,
which are automatically opens as a respond. By this
reason the smart environments support a bounded,
local context of user interaction. Smart
environments will also follow a novel and
revolutionary upgrades to be incorporated into the
environment in the sense of a support less obtrusive
interaction (pressure sensors can be for example Figure 4: Three models of ubiquitous computing:
incorporated into surfaces to detect a people sitting smart devices, smart environments and smart
location or walking over). interaction [1].

2.4 Adaptive Systems for Ubiquitous

Computing

Ubiquitous computing provides a vision of

computing systems which are located everywhere
around us, embedded in the things of our everyday
life. They provide an easy access to information
and communications bases dedicated to our current
location. People are able to interact with any
ubiquitous computing environment which they
attend. This is a reason why the ubiquitous
computing environments must respond dynamically
to specific user needs, resources dedicated to their
owner’s rights or to the current usage context.
Figure 3: Three models of ubiquitous computing: These require a high level of adaptivity which must
smart devices, smart environments and smart be provided by ubiquitous computing systems and
interaction [1]. related connecting networks [2].
Described project deals with several of issues
2.3 Smart Interaction related to providing such adaptivity for ubiquitous
computing environments which will be described
While a smart devices and smart environments more in the following sections.
support the core properties of UbiCom, an
additional type of design is needed to connect 3 REACTION ON A CHANGE OF
together their numerous particular activity LOCATION – LOCATION-AWARE
interactions. Smart interaction is needed to support ADAPTATION
interaction model between UbiCom applications
and their UbiCom infrastructure, physical world We can imagine the usage of such described
and human environments. In the smart interaction UAS in the information systems area of botanical or
design model, system components dynamically zoological gardens. In such areas there has been a
big potential of usage of a continual localization by by data prebuffering. Selecting of data objects to be
use of GPS or wireless networks (in case the GPS buffered to mobile device cache is made on the
has not a sufficient signal – e.g. in urban centres or base of position of user’s device. For every position
neighbourhoods with high buildings, forest parks or in area, where the prebuffering is being made, the
in deep valleys). There is also a possibility to location-aware objects for such user’s position
compute a current and predicted user track, so we exists. PDPT Core pushes a data from SQL
can predict a position of user in near future (e.g. 25 database (WLA database (Fig. 2)) to clients PDA
meters north in one minute). Usability of these on a base of PDPT Core decision algorithm.
information sources is uncountable.
One of possible use of user predicted position
is for a determination of a data, which will be
needed by user of mobile UAS in near future. Such
data (data artifacts) can be preloaded to user’s
device memory for future requests. The need of
preloaded artifacts grown from a need of up to date
data context of dynamic online system. Of course
when static offline system is used, there is a
possibility to load a needed data before usage (e.g
store artifacts at SD Card with a size limit to several
GB). When user request info about his location in
context of zoo or garden (turn-on the device is only
needed by user), the client application will respond
with a map of near surroundings and a prebuffered
data artifacts. User can select a documentary about
animals or vegetation around him which can be
viewed or played. User can act with direct requests
to selected kinds of these. These preferred kinds
will be taken into account to evaluate future
objects/artifacts and preloaded only the most
important ones for a user. The type of artifact is
also evaluable as well as his size because the user
may not want to look at too long or micro
presentation.
As client devices of online UAS, the mobile
wireless devices like PDA or Smart phones are
commonly used equipped with internet connectivity.
The connection speed of the two most common
standards GPRS and WiFi varies from hundreds of
kilobits to several megabits per second. In case of
online UAS or some other types of facility
management, zoological or botanical gardens,
libraries or museums information systems, the WiFi
infrastructure network is often used to interconnect
mobile device clients with a server. Unfortunately,
Figure 5: Scheme of WLA architecture PDPT
the low performance hardware components are used
server database.
in PDAs or SmartPhones due to a very limited
space. Due this a theoretical maximum connection
The benefit of using a PDPT consists in
speed is not reachable on such devices. The limited
reduction of time delay, which is needed to display
connection speed represents a problem for clients of
requested artifacts from PDA client. This delay
online system using large artifacts (data files). In
must not be longer than the time for which a user is
some specific cases it is not possible to preload
able to wait for some response from application.
these artifacts before the use of mobile device in a
Hence, the maximum response time of an
remote access state due any reason.
application (PDPT Client) for user must be
specified firstly. Nielsen in his book [8] specified
3.1 Low System Throughput on Current
this time delay to 10 seconds [9]. During this time
Mobile Devices
the user was focused on the application and was
willing to wait for an answer. The Nielsen book is a
The real downlink speed for WiFi network
basic literature for this phenomenon. Galletta,
(802.11b,g) is about 1280 kbit/s for modern PDA
Henry, McCoy and Polak (2002) findings suggest
devices [6], [7]. Primary dataflow can be increased
that, ‘decreases in performance and behavioral that the Visual Studio and classic SQL Server are
intentions begin to flatten when the delays extend to used. Small data amount for installation (2,5 MB) is
4 seconds or longer, and attitudes flatten when the also an advantage.
delays extend to 8 seconds or longer’. Based on this
knowledge, we defined this delay for our testing 4 REACTION ON A CHANGE OF
purposes to 5 seconds. For this time is possible to BIOMEDICAL DATA – ACTIVE
transfer (from server to client) a data amount of 800 CONTEXT-AWARE ADAPTATION
kB (for 1280 kbit/s downlink).
The next step was an average artifact size A key problem of context-aware systems
definition. The network architecture building plan design is to balance the degree of user control and
is used as a sample database, which contained 100 awareness of their environment. We can recognize
files of average size of 470 kB. The client two extreme borders as active and passive context-
application can download during the 5 second aware. In active context-aware system, the UAS is
period from 1 to 2 artifacts. The final result of aware of the environment context on behalf of the
several real tests and consequential calculations is user, automatically adjusting the system to the
definition of artifact size to average value of 500 context without the user being aware of it [1]. This
kB. The buffer size may differ from 50 to 100 MB is a useful in our application where a strict time
in case of 100 to 200 artifacts. constraints exists, because the user-patient cannot
due to immobility, or would not otherwise be able
3.2 Position Oriented Database to adapt to the context quickly enough. We are
using principles of UAS in area of biomedical data
If the mobile device knows the position of the processing, where we try to predict some kind of
stationary device (transmitter), it also knows that its problems by patient data analysis. We developed a
own position is within a range of this location context-aware Biotelemetrical Monitoring System
provider. The typical range varies from 30 to 100 m (BMS) as a part of the UAS and PDPT Framework
in WiFi case, respectively 50 m in BT case or 30 project facilitates the following:
km for GSM. Granularity of location can be • Real-time collection of the patient vital signs
improved by triangulation of two or more visible (e.g. ECG, EEG) by means of a Body Area
APs (Access Points) or using the more accurate Network (BAN) or direct wireless connection to
position algorithms (Monte Carlo localization). In PDA device monitoring station.
PDPT framework only the triangulation technique • Real-time transmission of the vital signs using
is used due to the sufficient granularity of user the wireless connectivity to the healthcare
position information. Monte Carlo localization was professionals through a complete architecture
tested in one segment of tested environment including a server database, web services, doctors
without marginal success (Time needed to web access to patients collected and preprocessed
implement algorithm was inadequate to position data.
quality results). Information about the user position • Seamless handover over different wireless
are stored in Position table (Fig. 2). Locator table communication technologies such as BlueTooth,
contain info about wireless AP with signal strength WiFi, GPRS or UMTS.
which are needed to determine user position. • Context-aware infrastructure to sense the
WiFi_AP, BT_AP and GSM_AP tables contain all context (e.g. location, availability, activity, role) of
necessary info about used wireless base stations. the patients and Emergency Response Services
WLA_data table contain data artifact along with (ERSs) to provide assistance to the patient in case
their position, priority and others metadata. of an emergency. An ERS could be fixed (e.g.
hospital) or mobile (e.g. caregiver). A mobile ERS
3.3 PDPT Client - Mobile Database Server is published in the BMS [18].

The large data artifacts from PDPT Server

(WLA_data table (Fig. 2)) are needed to be
presented for user on mobile device. In case of
classical online system the data artifacts are
downloaded on demand. In case of PDPT solution,
the artifacts are preloaded to mobile device cache
before user requests. As mobile cache the SQL
Server 2005 Mobile Edition was selected. Our
mobile cache contain only one data table Buffer.
Only the needed columns from PDPT server
WLA_data table were taken for mobile version
Buffer table. MS SQL Server 2005 Mobile Edition
was selected for easiest managing of them in case
 biosensors network (e.g. ECG, pulse, oxy, pressure).
The biomedical ECG data are continually processed
(in Real Time) through a complete infrastructure of
developed UAS. First false artifact recognition is
made on mobile measurement stations near the
patient to allow an immediate action from ERSs.
The more sophisticated data analysis is made at
server level. This data processing is made on the
base of neuron network and fuzzy logic behaviour.
Unfortunately, we reach only a small level of
successful false detection (patient problem
detection) up to date. In this area we are expected a
future impact of our solution. The low detection
rate is caused by several facts. Of course the better
algorithms are needed at the first, but this problem
cannot be solved satisfactory in near future.
Another problem is caused by a slow connection by
Figure 6: Flowchart of Reactive (Left – Fig.6.a) WiFi network, because some biomedical data
and Proactive (Right – Fig. 6.b) ERSs Selection and contain a huge amount of data. This problem is
Invocation Approach. possible to solve by our PDPT framework as a part
of our UAS solution. By this solving, we improve
Classical access to patients request are made by the quality of detection by a 40 % (median value of
reactive flowchart (Fig. 6.a), where a patient is 12 channels ECG). All the same, the real time
equipped with a classical offline measuring devices transfer rate is now still fail to reach.
with some type of alarms. Every violated alarm
need to be a carried out by doctor decision. Such 4.2 wakeNsmile Application – Proactive User
access is very time-consuming. Adaptive System
Second proposed access is based on a proactive
principle (Fig. 6.b), where the patient is equipped Proactive principle can be used not only in
with an online measuring devices with an online large distributed solution for medical centres, but it
connection to some kind of superior system (in our can be found usable in a many other solutions. One
case the BMS is presented). In this case, a patient’s of them we found in an application to allow for
measured data are processed on mobile monitoring people have a happy wake up. A mobile device
station or at server. An alert will invoke when the application was developed to solve a problem of
anomaly data are founded in patient’s records. bad wake up at morning for all the people.
Consequently the doctor is responsible to make a Sleep is a complex process regulated with our
decision to invoke other ERSs or to remove Alarm brain and as such is driven by 24 hour biological
(in case of false detection of anomaly). Such kind rhythm. Our biological clocks are controlled by
of behavior is based on UAS. In many of events a chemical substances that are mostly known to us.
predicted and solved problems can save a life. The Approximately two hours after we fall asleep
predicted patient’s problems are in most cases our eyes starts to move back and forth irregularly.
minor in compare to a major problems detected in Based on this fact scientists divided sleep stages
time where occurred. into two main stages REM sleep with (Rapid Eye
Movement) and NREM sleep stage (Non Rapid Eye
4.1 Biomedical Data Acquisition, Processing Movement). NREM sleep is divided into another
and Proactive Reaction four sub-stages, when with increasing number the
sleep is more and more deeper.
Our developed BMS can currently handle two During healthy individual sleep, REM and
types of biomedical data: NREM stages changes a few times. Most of the
1. 12 channels wireless ECG – BlueECG and dreams are happening in REM stage. Body muscles
2. Bipolar wireless ECG – corbelt. are completely loosened and thanks’ to this fact one
These data are measured, preprocessed on is awaken refreshed.
mobile monitoring station (PDA, embedded device, During deep (NREM 3 and 4) sleep stages
notebook), visualized on monitoring station’s blood pressure is decreasing which lowers chance
display (in available), sent by wireless connection of cardiovascular danger. Also growth hormone is
to web service and stored on server for produced in its maximum in adolescent age. Sleep
consequential access by doctors or medical personal. stages are possible to divide into several:
Used data acquisition devices provide a successful (a) Wake (Awake),
result in case of testing a developed solution. In (b) REM – we dream in this stage,
near future we plan to use a t-shirt with equipped (c) NREM1 – falling asleep,
(d) NREM2 – light sleep, network analysis and people database is planned for
(e) NREM3 – deep sleep, future steps to be a completely embedded solution
(f) NREM4 – deepest sleep. at Mobile UAS Framework. We executed several
wakeNsmile application (Fig. 7) was developed tests with very promising results. More than 70 %
in C# programming language and uses .NET of successful happy wake up of test persons at
compact framework version 3.5, which is a special morning without any restriction to test persons. For
derivative of .NET framework for mobile devices healthy test persons the results was more than 89 %
[7]. Application was developed in Visual Studio [13].
2008 Team Edition on Windows Mobile 6.5 Developed application act as a proactive
emulator and tested on a Hewlett-Packard mobile solution in sense of happy wake up of users in most
device (originally HTC Roadster) with Windows suitable body state. User adaptivity can be however
Mobile 6.5 operating system. Minimal requirements sustained by user’s inputs collection store (user
for application running are Mobile device with manual sets along with successfully detected mild
Windows Mobile 6 and higher and .NET compact sleep stages) to achieve a higher level of user
3.5 or higher. adaptivity based on them. Similar adaptation is used
wakeNsmile application uses user control by next project of intelligent alarm called “Gently
called Alarm, that has been created as a part of this Alarm” [15].
project. Application is using Math.NET neodym In wide context a limited type of green
library for FIR (Finite Impulse Response) filter pervasive computing system can grown from
design and WaveIn and WaveOut libraries for developed application as a base for such system
mobile device sound interface communication. inputs in form of users knowledge or users mental
state during a day after a successful or unsuccessful
happy wake up [16], [17].

5 REACTION ON A CHANGE OF LOGGED

USER – PERSONAL-AWARE
ADAPTATION

Next possible way to react on user needs is in

classical user input processing. Based on user login
a personal-aware adaptation of UAS can be defined.
Well known is a model of screen resolution
adaptation based on a used mobile device. Classical
way is in user setting module located in used
application. This however requests a user action at
each time a different user is logged in.

5.1 Adaptive User Interface for Mobile User

Adaptive System

To prevent such waste user time, user interface

adaptivity can be developed and used based only on
user login information. UAS server can collect a
Figure 7: wakeNsmile application example in user data such as a request of special user interface
Visual Studio 2008 Windows Mobile 6 emulator. layout (font size, buttons size and locations, wide of
scrollbars, etc.). After user login application is
wakeNsmile application is developed to react initiated in used best fitting scheme. Example of
on users declared request in form of happy wake up such user defined user interface is shown at (Fig. 8)
at predefined time (Fig. 7). The time defined for resp. (Fig. 9).
alarm is however the latest possible time to wake
up of user. We are trying to detect a body state in
which the user is most able to wake up with a smile.
Time period for detection analysis of state phases is
declared to 30 minutes. A Fast Fourier
Transformation (FFT) and some other sophisticated
methods are used for it. Created application is an
ideal example of user adaptive solution for mobile
devices. Currently a single application is developed,
but a distributed architecture version with a neural
 5.2 New Components for Mobile User
Adaptive Systems

However not every user is able to access small

fonts so user interface with a large elements of user
interface are welcome. Examples of such elements
are described in (Fig. 10). A first example presents
switchers (Fig. 10.a.). They provide a sizable
intuitive way to support an adaptation on user
ability. Every described element is developed as
components of UAS framework. Use in any other
projects is therefore very easy and comfortable.

Figure 8: User interface layout initiated based on

UAS server data. QVGA layout on a VGA display.

Figure 10: User interface components: 0/1 switch

(Upper-Left – Fig. 10.a), On/Off switch (Upper-
Right – Fig. 10.b) and navigation arrows where a
left direction is selected (Lower – Fig. 10.c)

Another component of UAS framework is

navigation arrows (Fig. 10.c.), which is a sizeable
component with one enumeration type of direction
which can be used to easily navigate in some
outdoor use cases.
Next component of UAS framework is circle
visualizer (Fig. 11.a.), which is a sizeable
component with two properties: color areas
definition and min-max values. This component can
be used to inform user about valued state of some
controlled properties in the context of their
boundary values. By use of this context a user can
get more complex information instead of classical
value information (e.g. in text/numerical form).
The last example of component is based on
previous circle visualizer component, which is
parent of a new component is sense of object
programming model. The component can represent
e.g. milliammeter (Fig. 11.b.) or voltmeter (Fig.
Figure 9: User interface layout initiated based on 11.c.) as a two examples of measurement
UAS server data. VGA layout on a VGA display. visualization component. From parent it inherits all
properties and it adds a text properties for type of
Depending on a user ability to view smaller meter which it is represent in real case. Of course
fonts an indispensable number of other rows are the shape is not a circle type, but it is rectangle.
viewable by user a higher resolution (Fig. 9).
Figure 11: User interface components of
measurement visualization: value of 17 at circle
visualizer (Upper-Left – Fig. 11.a), voltmeter
(Upper-Right – Fig. 11.b) and milliammeter (Lower
– Fig. 11.c).

More information about user adaptation

improvement using software components for
mobile control systems in .NET Compact Figure 12: User Adaptive System Framework
Framework can be found in [14]. architecture.

6 THE USER ADAPTIVE SYSTEM Artifact data object can be defined as a

FRAMEWORK multimedia file type in complex-awareness, which
represent an object in Position Oriented Database –
A combination of a predicted user position with table WLA_data with time, position and
prebuffering of data, which are associated with biomedical-awareness. To manage locations of
physical location bears many advantages in artifacts, firstly the building map is needed. The
increasing throughput of mobile devices. The key position of corporate APs is also needed to
advantage of PDPT solution in compare to existing determine a user position based on a distance from
solutions is that the location processing, track each visible APs. All obtained positions info need
prediction and cache content management are to be stored in UAS server database through a
situated at server side. The solution allows for PDPT Core web service. Artifacts with position
managing many important parameters (e.g. AP info coordinates are stored in WLA_data table by use of
changes, position determination mechanism tuning, “WLA Database Artifact Manager”. This software
artifacts selection evaluation tuning, etc.) online at application was created to manage the artifacts in
a server. By adding a Biomedical Data Processing Position Oriented Database.
solution, the Complex User Adaptive System The PDPT prebuffering principle consists of
(UAS) Framework is growing from (Fig. 12). several following steps:
While the whole PDPT Framework concept allow 1. Client must activate the PDPT buffering
to manage a artifacts in context-awareness and checkbox on PDPT tab at PDPT Client, which
time-awareness, the UAS Framework shift these creates a list of artifacts (PDA buffer view sample
possibilities to manage artifacts in biomedical which contain only ID of artifacts), which are
context-awareness allowing a response for example contained in his mobile SQL Server CE database.
to user´s non declared needs. 2. PDPT Framework Core web service module
Biomedical Data Processing sensor at Mobile creates own list of artifacts (imaginary view sample
Device side of architecture (Fig. 12) collect of PDA buffer) dedicated to actual user device
information from user’s body through a Bluetooth position. It also compares it with real “PDA buffer
connection to any kind of wearable biotelemetrical view sample”. The area is defined as a 3D rectangle
devices. These data are transferred to UAS Server object where the user’s position is located in center.
along with locator module data, which is processing 3. The PDPT Core continues in next step with
these knowledge to act with adequate reaction in comparing of both images. If there are some missed
sense of user comfort improvement as a response artifacts in PDA buffer, they are prebuffered to
time reducing for requested information by data PDA buffer. When all artifacts for current user
prebuffering or any other reaction (e.g. screen position are prebuffered in PDA buffer, there is no
resolution improvement, display brightness etc.). difference between images.
4. After all artifacts are prebuffered to PDA buffer prebuffering is needed to be defined and also the
of mobile client, the PDPT Core is going to make data artifacts must be defined. One way is in use of
steps 1 to 3 once more for a new predicted user developed software “WLA Database Artifact
position with new enlarged area (3D rectangle). Manager” for offline case, but the useful solution is
in determination of large data objects in online case.
7 DISCUSSION OF RESULTS Such determination is not easy. Possible solution
can be seen in application of Position Oriented
The PDPT Framework project is developed Database scheme to convert an existing server
from 2005 until now in several consequential database of online system to Position Oriented
phases. Current state of the project is near the real Database structure. After such conversion, the data
company stocking. Final tests were executed in are possible to select based on position in stocking
university campus of Technical university of area. Consequently if data object – artifact can be
Ostrava. For company stocking is possible to think selected, the PDPT server can prebuffer such data
about several areas. These possibilities will be to mobile device.
discussed in [section VII.2]. As a summary, the PDPT is now usable at
immobile patients at 100% successful rate of
7.1 Final Test Results of PDPT Framework prebuffered artifacts. Such sort of patients is
Part specific for only low speed of their transfer inside
the environment. Due this fact the PDPT is
For testing purpose, five mobile devices were functioning. If the environment for prebuffering
selected with different hardware and software will be equipped with a higher number of WiFi APs,
capabilities. Six types of tests batches were the usability of PDPT in dynamic cases will be
executed in test environment. Two different test much more achievable.
scenarios were executed as static and dynamic tests
scenarios. Static test was based on a predefined 8 CONCLUSIONS
collection of data artifact which belongs to defined
user position in test environment. Five test position A concept of UAS as well as PDPT Framework
were selected where approximately 12 data artifacts and BMS Framework was described with main
was needed to successful prebuffering. Three focus on Position Oriented Database on server and
iterations were repeated in each position. If any of mobile devices. Coexistence of proposed solutions
these expected artifacts stay un-buffered, the is in unnumbered areas and the results of complex
quality of prebuffering is going low. The tests were solution are better than expected. Also the final
performed with result from 69,23 % to 100 %. The static and dynamic tests were performed and
mean value of test results was 93,63 %. From all 15 discussed. The developed UAS can be stocked on a
tests, the 9 were executed with a 100 % of wide range of wireless mobile devices for its main
successful score. issue at increased downlink speed. The localization
Every dynamic test was between two points part of PDPT framework is currently used in
with 132 meter distance. Every even test was in another project of biotelemetrical system for home
reversed direction. Five iterations (five devices care agencies to make a patient’s life safer. Several
used) were made during one batch. Results provide areas for PDPT stocking was founded in projects of
a good level of usability when user is moving Biotelemetry Homecare. In these selected areas the
slowly (less than 0,5 m/s). This fact is caused by use of PDPT framework is not only partial, but
low number of visible WiFi APs in test complete include the use of wide spectrum of
environment, where 60 % of time only 1 AP was wireless communication networks and GPS for
visible, 20 % for 2 visible and 5 % for 3 or more tracking people and urgent need of a high data
visible WiFi APs. 15 % of time represents a time throughput on mobile wireless connected
without any WiFi connections. Reached values of monitoring devices. Several of UAS principles can
prebuffering quality in such case are very good. be used there also. These possibilities will be
investigated in future.
7.2 Possibilities of Using a PDPT Framework
in Real Environment ACKNOWLEDGMENT

Dynamic tests of PDPT Framework show the This research has been carried out under the
problem of a low number of visible WiFi APs for financial support of the research grants “Centre for
localization determination in the test environment Applied Cybernetics“, Ministry of Education of the
of university campus. For the real case of usage and Czech Republic under Project 1M0567 and “Safety
for the high level of prebuffering quality, the and security of networked embedded system
minimal number of simultaneously visible WiFi applications”, GACR, GA 102/08/1429, Grant
APs at each place of stocking area must be from 3 Agency of Czech Republic.
APs. For successful stocking of PDPT, the area of
9 REFERENCES 20040030832, (2004)
[12]Krejcar, O., “Using of Ubiquitous Computing
[1] Poslad, S.: Ubiquitous Computing: Smart Principles to Develop a Mobile User Adaptive
Devices, Environments and Interactions, John System Framework”, In Proceedings of 9th
Wiley & Sons, Ltd, London,UK, ISBN 978-0-470- RoEduNet IEEE International Conference,
03560-3 (2009) RoeduNet 2010, 24. – 26. June 2010, Sibiu,
[2] Lewis, D., O'Sullivan, D., “Adaptive Systems Romania, pp. 352-357, ISBN 978-1-4244-7335-9,
for Ubiquitous Computing”, In Proceedings of the ISSN 2068-1046, (2010)
1st international symposium on Information and [13]Krejcar, O., Jirka, J., Janckulik, D., “Proactive
communication technologies, ACM International User Adaptive System for Windows Mobile
Conference Proceeding Series; vol. 49, pp. 156, Devices – Processing of Sound Input Signal for
(2003) Sleep State Detection”, in Proceedings of 2nd
[3] Satyanarayanan, M. “Pervasive computing: International Conference on Mechanical and
vision and challenges”. In IEEE Personal Electronics Engineering (ICMEE 2010), vol. 1, pp.
Communications, vol 8, pp. 10–17. (2001) 374-378, Kyoto, Japan, August 2010. DOI
[4] Coen, M. H., “Design principles for inteligent 10.1109/ICMEE.2010.5558525
environments”. In Proceedings of 15 th National / [14]Krejcar, O., Cajka, J., User Adaptation
10 th Conference on Artificial Intelligence / Improvement Using a Software Components for
Innovative Applications of Artificial Intelligence, Mobile Control Systems in .NET Compact
pp. 547–554. (1998) Framework, In Proceedings of International
[5] Cook, D.J. and Das, S.K., “How smart are our Conference On Networking and Information
environments? An updated look at the state of the Technology, ICNIT 2010, 11. – 13. June 2010,
art.” In Pervasive and Mobile Computing, 3(2): pp. Manila, Philippines, NJ. IEEE Conference
53–73. (2007) Publishing Services, 2010 pp. 545-549, ISBN 978-
[6] Krejcar, O., Prebuffering as a way to exceed 1-4244-7579-7, DOI 10.1109/ICNIT.2010.5508453
the data transfer speed limits in mobile control [15]Gentle Alarm iPhone Application developed by
systems, In ICINCO 2008, 5th International Craft mobile company (http://gentle-alarm.com/)
Conference on Informatics in Control, Automation 2010
and Robotics, May 11-15, 2008 Funchal, Portugal, [16]N. Chilamkurti, S. Zeadally, S. Jamalipour, and
pp. 111-114, (2008) S. K. Das, “Enabling Wireless Technologies for
[7] Krejcar, O.: Problem Solving of Low Data Green Pervasive Computing,” in EURASIP Journal
Throughput on Mobile Devices by Artefacts on Wireless Communications and Networking, vol.
Prebuffering, In EURASIP Journal on Wireless 2009, Article ID 230912, 2 pages, 2009.
Communications and Networking, Article ID doi:10.1155/2009/230912
802523, 8 pages. Hindawi publishing corp., New [17]N. Chilamkurti, S. Zeadally, and F. Mentiplay,
York, USA, (2009) “Green Networking for Major Components of
[8] Nielsen, J.: Usability Engineering, Morgan Information Communication Technology Systems,”
Kaufmann, San Francisco, (1994) in EURASIP Journal on Wireless Communications
[9] Haklay, M., Zafiri, A.: Usability engineering and Networking, vol. 2009, Article ID 656785, 7
for GIS: learning from a screenshot. The pages, 2009. doi:10.1155/2009/656785
Cartographic Journal 45(2), 87–97 (2008) [18]Pawar, Pravin and Beijnum van, Bert-Jan and
[10]Brasche, G. P., Fesl, R., Manousek, W., Salmre, Mei, Hailiang and Hermens, Hermie (2009)
I. W., “Location-based caching for mobile devices”, Towards Proactive Context-Aware Service
United States Patent, Microsoft Corporation Selection in the Geographically Distributed Remote
(Redmond, WA, US), 20070219708 (2007) Patient Monitoring System. In: 4th International
[11]Squibbs, R. F., “Cache management in a Symposium on Wireless Pervasive Computing,
mobile device”, United States Patent, Hewlett- ISWPC 2009, 11-13 Feb. 2009, Melbourne,
Packard Development Company, L.P., Australia.