Intelligent Context-Aware and Adaptive Interface for Mobile LBS

Abstract

Context-aware user interface plays an important role in many human-computer Interaction tasks of location based services. Although spatial models for context-aware systems have been studied extensively, how to locate specific spatial information for users is still not well resolved, which is important in the mobile environment where location based services users are impeded by device limitations. Better context-aware human-computer interaction models of mobile location based services are needed not just to predict performance outcomes, such as whether people will be able to find the information needed to complete a human-computer interaction task, but to understand human processes that interact in spatial query, which will in turn inform the detailed design of better user interfaces in mobile location based services. In this study, a context-aware adaptive model for mobile location based services interface is proposed, which contains three major sections: purpose, adjustment, and adaptation. Based on this model we try to describe the process of user operation and interface adaptation clearly through the dynamic interaction between users and the interface. Then we show how the model applies users' demands in a complicated environment and suggested the feasibility by the experimental results.

Full text

Research Article
Intelligent Context-Aware and Adaptive Interface for
Mobile LBS
Jiangfan Feng1,2 and Yanhong Liu1
1College of Computer Science and Technology, Chongqing University of Posts and Telecommunications, Chongqing 400065, China
2Key Laboratory of Instrument Science and Dynamic Test, North University of China, Taiyuan 030051, China
Correspondence should be addressed to Jiangfan Feng; fengjf@cqupt.edu.cn
Received  November ; Revised  January ; Accepted  February 
Academic Editor: Dongrong Xu
Copyright ©  J. Feng and Y. Liu. is is an open access article distributed under the Creative Commons Attribution License,
which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Context-aware user interface plays an important role in many human-computer Interaction tasks of location based services.
Although spatial models for context-aware systems have been studied extensively, how to locate specic spatial information for
users is still not well resolved, which is important in the mobile environment where location based services users are impeded
by device limitations. Better context-aware human-computer interaction models of mobile location based services are needed not
just to predict performance outcomes, such as whether people will be able to nd the information needed to complete a human-
computer interaction task, but to understand human processes that interact in spatial query, which will in turn inform the detailed
design of better user interfaces in mobile location based services. In this study, a context-aware adaptive model for mobile location
based services interface is proposed, which contains three major sections: purpose, adjustment, andadaptation. Based on this model
we try to describe the process of user operation and interface adaptation clearly through the dynamic interaction between users
and the interface. en we show how the model applies users’ demands in a complicated environment and suggested the feasibility
by the experimental results.
1. Introduction
USAScholarSchlitrstproposedthreetargetlocation
based services (LBS) in : spatial information, social
information, and resources nearby. Nowadays, mobile LBS
interface has signicant inuence, while a mass of users
carry their mobile terminals to require location services
such as the wish to nd a better route all over. Under this
circumstance, problems of traditional mobile LBS interface
are revealed which lack of ability to adapt users’ demands
initiative. Context-aware human-computer interface makes
mobile LBS interaction more natural and ecient which is
able to adapt to dierent users’ characters and requirements
bytakingadvantageofusableinformationaboutusers’tasks
(e.g., locations and preferences of users, experiences, and sur-
rounding environments). Many of these adaptive interfaces
serve specic users, such as the interface which Cao designed
for children that apply cartoon-icon in  []andtheuser
interface adaptation proposed by Zouhaier et al. which is
based on context awareness for disabled people in  [],
for the characters of specic users are obvious. And these
researches are based on the human-computer interaction
model which describes the characteristics of the interaction
process between human and machine. Some early researches
of adaptive interface model are based on user features,
like Rich proposed a modeling method which classies the
users based on their background and then provide dierent
services []. User modeling concentrated not only on users’
cognitive or reason, such as knowledge, goals, and planning
[],butalsoonemotionalandpersonality[,]. In the eld
of mobile LBS, Shi and Bian developed an adaptive expression
of spatial information and the adaptation policy of the
interaction elements on LBS interface in  []andLathia
et al. proposed state-of-the-art advanced traveler information
system (ATIS) which can adapt to users’ environment and
activities in  []. However, there are certain problems and
shortcomings in the current study.
On the one hand, most of context-aware adaptive inter-
faces are designed for specic tasks or applications, and many
researchers construct dierent models to satisfy users’ various
Hindawi Publishing Corporation
Computational Intelligence and Neuroscience
Volume 2015, Article ID 489793, 10 pages
http://dx.doi.org/10.1155/2015/489793

 Computational Intelligence and Neuroscience
requirements that are ignoring the diversity of one person.
Classication for users in one respect cannot represent the
features on their other aspects. On the other hand, context-
aware information is always used to predict the usable inter-
face but ignores its eect in the dynamic interaction process.
We maybe consider other aspects such as preference when
we recommend a suitable interface to a user built on his/her
cognitive ability. e challenge of adaptive interface in mobile
LBS is not simply to provide users information whenever
and wherever but also to provide appropriate information for
users when they need. e current research on the interface
adaptation lacks the exploration of user dynamic interactive
behavior. When the passage of time, task, and context
information change, then the content of adaptation changes.
According to these problems, this paper proposed dynamic
adaptive model and presents a corresponding method.
In view of the above problems, this paper presents a
context-aware adaptive interface for mobile LBS. At rst, we
establish a user model which has better generalization and
dierentiation degree based on users’ basic characteristic and
thebehaviorcharacteristic,andthenwematchtheusermodel
with the rene interface element modules; proposed adaptive
interface modeling method and system structure combine
with dynamic interaction behavior. At last, we explain the
adaptive process through a scenario. In the following Related
Works section the feasibility and applicability of context-
aware interface to be adapted to users in some solutions are
discussed.
2. Related Works
issectionshowsthefocusthatwedescribedbelowwithin
context-aware adaptive interface for mobile LBS existing
literature. ere are three parts that we discussed: context-
aware technology, adaptive user interface, and adaptive spa-
tial information.
2.1. Context-Aware Information. People oen naturally used
implicit information to make the content rich when there is a
process of human to human interaction for they understand
the situation of each other while for computers it seems dif-
cult to master this skill in comparison. erefore, context-
aware technology was used widely in order to attain the
purpose of natural interaction. Context is determined by
Merriam-Webster’s collegiate dictionary as “the interrelated
conditions in which something exists or occurs.” To put
it more specically, Schilit and eimer []proposedthat
context contains location and identities of nearby people
and objects in , and in  Brown et al. [] added
time of the day, season of the year, and temperature to the
original denition. Up to now, context is broadening to a
comprehensive concept including task context, user context,
and circumstance context. Generally speaking, context based
on mobile phones can be divided into three parts as follows
[]:
() user environment: location, preference, experiment,
social relations, and so forth;
() mobile environment: device suitable for users to input
or display, network, Bluetooth, and so forth;
() physical environment: weather, date, noisy, and so
forth.
Moreover, context-aware technology has the capability to
sense, detect, and grab the environment around users and get
the dynamic changes to speculate their behavior [].
Context-aware technology plays an important role in
mobile terminals which equip a rich set of sensors (e.g.,
camera, accelerometers, GPS, digital compass, gyroscope,
ambient light sensors, proximity sensors, multitouch panels,
and microphone) []; it also enriches the function of GIS to
provide users a variety of services. Tomitsch et al. discussed
the context of human actions in public space and how they
fed back []. Lathia et al. []proposedmobiletraveler
information system which can become personalized services
based on explicit preferences. J. Karat and C.-M. Karat
[] proposed context-aware route recognition approach to
improve the accuracy of routing recognition. Abowd et al.
proposed a mobile context-aware tour guide in  [];
Cai species a semantic model which combined with context
and demonstrates how this model supports contextualized
interpretation of vague spatial concepts during human-GIS
interactions in  []. Chung and Schmandt proposed
a mobile user-aware route planner which can learn a user’s
everyday routes and provides directions from locations along
thoseroutesin[].
2.2. Mobile Adaptive Interface. Human-computer interface
(HCI), which is also known as the user interface, is media for
the exchange of information between user and computer. e
traditional design methods consider the eciency problem
of using rarely, and the traditional interface can only adapt
to a few people, but also cannot meet the requirements for
one person in dierent periods with the xed user interface
designed according to users’ average level while the computer
used popular and user group became more and more widely
used. Adaptive user interface (AUI) which can adjust itself to
t a user or a task [] emerged and developed fast while the
requirement of omnipresent computing challenge traditional
interfaceemergedandincreased.Earlierintheresearchof
an adaptive user interface, it requires three models: system
model, user model, and the interaction model []. e
system model describes the characteristics of the system
that can be changed, such as the system to be able to
adaptive. Acquisition and application of the user model are
thefoundationofanadaptiveuserinterfacetomakethe
system adapt to the individual user behavior. Interaction
model denes how the system is modied, and what it can
adaptto.Aboveall,thedegreeofadaptabilityintheadaptive
process depends on the user model which describes users’
knowledge that can be utilized to facilitate human-computer
interaction.
Many researches take advantage of adaptation to provide
personalized services for users such as helping users to
obtain information, giving users a recommendation, tailoring
information for users, or providing help. Yoon et al. proposed
an adaptive mixture-of-experts model to solve the complexity

Computational Intelligence and Neuroscience 
and personality problem of multiuser interface in  [].
AndChengandLiudevelopedanadaptiverecommendation
system that inferred users’ preferences and adjusted the user
interfaces []. Wang et al. presented an automatic approach
which helps users who suer from visual impairment to make
use of online map with independent access to geographic
direction []. Sulaiman and Sohaimi []discussapossible
interface which is simple enough for older users through
analyzing the situation of using a mobile phone.
2.3. Adaptive Spatial Information. Nowadays, people can
carry mobile devices everywhere and every time with the
development and extensive application of mobile communi-
cation and internet technology. In this case, a large amount
of requirements concentrated on the interests of users them-
selves: environment information such as recommending
interest point to users by acquiring the users’ location.
We oen need to consider the personal issues for spatial
information used by more and more users. In particular,
the users which have mobile phones with dierent running
speed needed dierent degrees of information presentation,
and the users with unique moving speed needed dierent
scale. In addition, dierent users request dierent aspects
of spatial information; for example, tourists pay attention
to scenic spots while drivers follow with road conditions.
Another adaptive problem is how to determine quantity
of information displayed on the screen. Many navigation
charts are based on accessibility to display all the details
when facing the problem. In fact, showing the map too
detailed not only is dicult to understand and display but
also makes the user focus on useless information which limits
the eectiveness. erefore, mobile adaptive visualization of
spatial information must be based on users’ needs to provide
details step by step [] such as providing user detailed
information when he/she amplies the map gradually.
In summary, the key elements of adaptive spatial infor-
mation are related with the user, the mobile terminal, and
the environment. User aspects include user background
(such as physiological dierences, preference dierences,
and cognitive dierences), user location (position, speed,
and direction), and user requirements. Environment aspects
include basic information (such as temperature, weather,
date, and time), and related users’ information means the
information of the other users which related to the users’
tasks. And the eects of system contain transmission speed,
information receiving rate, and so on.
Adaptive interface has been extensively studied, but the
self-learning user interaction lacks. Here we adopt the use of
adaptive user interface model to predict user’s intent eec-
tively with their spatial experience. Moreover, the proposed
use of experience awareness assists in the prediction of user’s
intent while satisfying the early re detection requirement.
3. Adaptive User Interface Model
Cognitive psychology regards people as an information
processing system, and people oen have dierent action
according to the environment, cognition, and personality
tendency dierences to reach the established goals [].
ere are interaction spaces between human, computer, and
environment. If the information presented on the interface
can adapt to the user’s cognitive psychology and personality
traits, users will complete the task quickly for reducing the
user operations.
Adaptive user interface model consists of the following
components: user model (UM), task model (TM), interaction
model (IM), domain model (DM), environment model (EM),
and presentation model (PM).
3.1. User Model. e user model needs to abstract the indi-
vidual dierences of users which may relate with personalized
service for no two users are identical. e interface style may
be aected by personality or preference dierences while the
expression mode of interface may be aected by cognitive
or physiological dierences. Here we dene the user model
(UM) as a collection: UM = {User ID, Knowledge, Physiology,
Inclination}.User ID represents a unique identier of user.
Andwedivideusers’backgroundintothreeparts:Knowl-
edge,Physiology,andInclination.Knowledge summarizes the
knowledge level of the user, Physiology is on behalf of users’
physiological characteristics, and Inclination represents the
subjective desire in every aspects.
At rst, people always associate academic quantication
when it comes to knowledge, but we use the concept to
represent users’ skill level including prociency in inter-
face using and cognitive ability on the map. Knowledge =
{Education, Professional level, Prociency level}.Andwecan
see the dierences existed between expert users in a eld on
soware prociency through the description. Second, factors
of physical abilities cannot be ignored in the interaction.
Physiology = {Age, Sex, Health}.Andlast,Inclination=
{Occupation,Personality,Habit,Preference};thesefactorscan
inuence the choice tendency of users.
3.2. Interface Static Elements. e task model was divided
into abstract task model and specic task model. Catch and
abstract the users’ needs and described the needs as abstract
tasks. Describe the interactive behavior in the system and
the dynamic behavior in the process of the interaction. e
task is an activity which is used in order to satisfy the user’s
goals. We can abstract task as TM = {Operation, Object, TC}.
Operation means a task operation which act on an object.
e object means an object which needs to operate. TC
represents the task context. e task model can be embodied
into STM = {tID, operationType, dataItem, dataType, C}.tID
means the task ID which needs to complete. operationType
is the type of interaction operation such as read, write, or
command.dataItemisadataitemwhichconsistsofdataID,
data attributes, and data value. dataType means the data type
of the operate data. 𝐶is a set of constraints to the data item.
Specify the task context to the constraint of the data and data
manipulation.
e domain model (DM) is dened as DM = {Object,
Attribute, Contact, Time};Objectisasetofinterfaceobject
at a specic time. e interface object is changing when the
environment changes or user task changes; therefore we use

 Computational Intelligence and Neuroscience
Time which is corresponding with the Time of TM and EC to
distinguish.
e interface model (IM) describes the interface, and
express the various controls manipulation in the dynamic
interactive process. Adjust the user interface components
and structure according to the specic task which is ana-
lyzedthroughthetaskmodel.einterfacemodelcanbe
abstracted as IM = {Control, controlConstraint, controlRe-
lation}. Control contains control ID and control attribute.
controlConstraint means the constraints of the controls on
the interface. controlRelation means the relationship of the
controls on the interface.
e presentation model (PM) is dened as PM = {Module,
MEC, MUC, MTC}; Module is a collection of interface
components. MEC, MUC, and MTC are component proper-
ties under the inuence of the environmental context, user
context, and task context.
3.3. Interactive Context. We can divide interactive context
(IC) into three parts: environment context (EC), user context
(UC), and task context (TC). Interactive environment con-
text can be divided into user environment and equipment
environment, where the user environment includes time,
place, and weather and equipment environment includes
transmission rate, and resolution. EC includes the user envi-
ronment and the device environment. e user environment
includes perimeter environment which may aect operations
of the user and user’s own context environment. Device
environment includes transmission rate and resolution. EC =
{DE, UE}; DE means the device environment and UE means
the user environment. DE = {MS, SO, TR}.MSmeansthe
movement speed of the device; SO is the screen orientation,
andTRistransmissionrateofthedevice.UE={Loc, Sur,
Soc, Act}. Loc means the location information of the user. Sur
represents the surrounding users of the user. Soc means the
social information of the user which can be obtained from the
social soware open interface. Act is the information which
is provided by the previous operation.
3.4. Interaction Methods. A large amount of contextual infor-
mation and the changeable situation requires a decision
mechanism to determine what kind of context information is
needed and when the information can be used. e decision
mechanism has three points which must be paid attention
to: selecting the appropriate context, allocating context pri-
ority level reasonable, and having the dynamic adaptability
(environment-stimulation, task-stimulation, etc.).
e most important in the interaction process is to
understandthepurpose,andthenwemustreducethe
problems caused in the process of achieving this goal; the last
is to let the user interface elements t the user like Figure .
e interaction model adjusts interface mode constantly
when some factors dynamic changes and stimulates.
4. Interaction Strategy
e adaptive user interface framework which based on the
GIS interface model is presented in the Figure .eFigure
describes the process of context information collection and
theeventualcapture,andexplainshowtoanalysisandhandle
events through the certain reasoning mechanism. According
to psychology related research achievements, people will
divide continuous events into several activities in perceptual
according to kinds of characteristics. Individual dierences
in cognitive exibility may underlie a variety of dierent user
behaviors [].Andusersinthesameactivitiestendtorepeat
steps which can reduce the user experience.
Combine the knowledge base to identify the user’s inter-
action patterns and predict the most likely interaction behav-
ior candidates of the user; then the adaptive recommendation
resultsappearedontheinterfacelayer.eadaptiveuser
interface framework interconnected between layer and layer;
the model layer determines the need of context information
collection, the adaptive layer is used to realize the user
interface adaptive function and scheduling, and the interface
layer is used to present the results of self-adaptation.
e adaptive decision-making mechanism is imple-
mented by capturing the user interaction sequence. When the
same user completes a task on the mobile GIS interface, the
same action sequence is oen repeated. So we can predict
the future behavior of the user through the interaction
sequence judgment, storage, and matching. e tasks which
were completed by the GIS human-computer interface can
be rened, and the user operation of every subtask will have
certain regularity which also contains its unique personalized
information. User actions can be rened into lower levels
of atomic operations, such as a button click, an operation
of input box, and the map zoom event. We can describe a
user action as A = (action object, action type). Several user
actions compose an action sequence, which can be collected
and matched to predict the most likely next step of users.
en adjust the interface elements dynamically and achieve
the goal of continuously reducing the user operation.
In the process of matching the sequence, we judge the
next action according to the front action, but the longer the
length of the sequence matching, oen the better the results.
erefore,weshouldchoosemoresuitablelengthsequenceto
match. Dening the average length probability of the match
sequence is 𝐿(𝑎|𝑠)=𝑙
𝑡(𝑎,𝑠)/∑𝑖𝑙𝑡(𝑎𝑖,𝑠). In addition, the
happeningoftheactionmayalsoberelatedtotheotheraction
andnotjustrelatedtothematchingofmodellength,soweuse
frequency of action occurrence 𝑃(𝑎 | 𝑠) = 𝑓(𝑎,𝑠)/∑𝑖𝑓(𝑎𝑖,𝑠)
to describe the action occurrence probability. We use the
action prediction 𝑅𝑡(𝑎,𝑠) = 𝐿(𝑎 | 𝑠) + 𝑃(𝑎 | 𝑠) to determine
the action occurring possibility.
e interactive action sequences appearing occasion-
ally can be seen as preinteraction pattern, which occurred
repeatedly will be put into the pattern library. e current
interaction sequence is matching with the action sequences
library, and the matching starts from the current action and
increases in length gradually under the context environment.
Obtain all the forecast candidate set; then choose the action
which has highest action evaluation 𝑅max as the prediction
results like Figure .𝐸means the new action set and 𝐶means
the existing action set.

Computational Intelligence and Neuroscience 
Environment Environment
Interface
elements
Style
Professional
Prociency
Domain knowledge
User
Decision mechanism
Interface
···
F : e environment of interaction.
Purpose
Basic
Better
TM
EC
More context
Basic context
Task context
Change
Change
F : User interface adaptive elements.
5. The Modeling Method
In the third part of the paper the adaptive user interface
model is put forward and this section will illustrate the
construction process from the abstract model to the specic
model. Create an adaptive interface model on the basis of the
user model, extract the information from the user model to
form the domain model, and then extract the management
tasks in the eld of domain model to form the task model.
e rst step is to build a user model. e user ID here
refers to the account of each user in the system which is
used to record dierent user information. Knowledge here
means the prociency of the user used navigation soware
andprofessionalleveloftheuser(Figure). Dierent levels
of users can lead to dierent operations. And the Physiology
refers to the aspects of users’ age and gender dierences.
We will comprehensive considering these aspects in the
experimental personnel selection. Inclination information is
recorded automatically in the system.
Domain model which is considered from the user model
needs to list managerial entity objects and analyze the
properties of these entity objects. For example, when a user
nds the route, entities in the domain include buttons, input
box, and map, and the interface elements of other services
may be involved for dierent users. e relationship between
the interface elements is the ordinal relation in the operating
sequence which is described in the adaptive strategies.
Each interact action corresponds to a task in the task
model, such as the switch interface, input box operation, and
determining button click. For example, a user wanted to nd
the point of restaurant; the description of the scenario in the
user model can be dened as US =⟨(UserID1,Prefer-
ence),(Device1,User En1)⟩,UserEn1=⟨Location,restaur-
ants,Pre-operation⟩;thecorrespondingtaskmodelcanbe
dened as Task =⟨AT1,AT 2, AT3⟩,AT1=⟨Click,Button-
Start,TC1⟩,AT2=⟨Click,ButtonSelect,TC2⟩,AT3=
⟨Zoom,Map,TC3⟩,TC=⟨Pre-operation,OPtime,Null⟩.
e building of interaction model according to each task
of the set of events in the task model described the atomic
operations in the user interface such as clicking on and
long press and described the corresponding commands of
interacting objects, such as a jump and zooming.
6. Experiment and Analysis
6.1. e Scene. In order to verify the result of the study, the
following scenario is designed to verify that the adaptive
methodiseasytouse.Andthenweevaluatesomevalues
which are measurable.
A Traditional Route and POI Searching Is as Follow
() User enters the application.
() Gototheweatherpagetochecktheweather.
() Go back to the main page.
() Click the button to enter the route searching page.
() Enter the locations to which the user wants to go.
() Clickthebuttontochoosetransportation.
() Click the button to display the route.
() Complete the route lookup.
() Click the button to enter the POI selection page.
() Choose a specic point of interest.
() e points of interest appeared on the map.

 Computational Intelligence and Neuroscience
Model denition
UM
DM
TM
PM
IM
Adaptive strategy
Context collection
EC
TC
UC
Decision mechanism
Stimulation User intent
Context
adjust
Context
analysis
Adaptive interface
Adaptive
assessment
Adaptive
algorithm
F : Mainly adaptive system.
Match
Take the length of Action evaluation
set
Take the last
action
Take the action of
highest R
Calculate
Jude
exhausted
Build
Preinteraction
pattern
Match the
interactive mode
Update pattern
library
Update C
Lsequence S
a∈E
a∈C
F : Sequence matching.
() Choose one of them to show the route.
() Complete the POI searching.
A Traditional Map Operation Is as Follows
() User logs into the application.
() User clicks the search button.
() Enter the search site in the input box.
() Click the conrm button.
() Zoomintocheckthelocation
() Zoom in once again.
() Click on the map sign to check the specic location
name.
Context-Aware Adaptive Interface for Route and POI Searching
Is as Follows
() User clicks the trac mode button (preferences
record) to enter the application.
() Choose the trac mode (advice according to the
weather condition) and enter the application.

Computational Intelligence and Neuroscience 
T : Context-aware adaptive interface scene design.
User System
() User clicks the trac mode button. () Display “bad weather”; recommend another trac mode.
() User chooses trac mode and clicks the button. () Display the location input box.
() User inputs the location and clicks the OK button. () Show the route and the interest point button.
() User clicks the interest point button. () Display the points of interest.
() User clicks the interest point. () Display the route.
() User logs into the application. () Record the user’s operation.
() User clicks the search button. () Enter the search interface.
() Enter the search site in the input box. () Operation sequence matching and display location markers directly.
() Zoom in to check the location. () Operation sequence matching and zoom in again.
() Click the map sign. () Display the specic location name.
() Enter the locations to which the user wants to go.
() Click the button to display the route.
() Complete the route lookup.
() Click the button (preferences record) to show the
points of interest.
() Choose one of them to show the route.
() Complete the POI searching.
Context-Aware Adaptive Interface for Map Operation Is as
Follows
() User logs into the application.
() User clicks the search button.
() Enter the search site in the input box.
() Zoomintocheckthelocation
() Click on the map sign to check the specic location
name.
Analysis of these scenarios can be found that, the tradi-
tional routes and POI searching need more options, for more
user active choices, which will produce more returns and
select operation. Here is the design of context-aware adaptive
interface which is user centered (Table ).
Figure shows the route searching interface. e appli-
cation icon is displayed as trac mode according to user’s
preferences. It will prompt the weather condition when user
clicksthebuttonandgiveadvice.Clickontheicontoenter
the location input interface. e context-aware adaption can
reduce the trac mode selection operation and also give
advice according to the environment actively.
Figure shows the adaption of interesting point search-
ing. In the process of user walking, nding interesting point
and giving corresponding button to users according to the
preferences can meet users’ requirements more easily. Click
the button to access the route.
Figure shows that the system recorded the user action
sequence and forecasted the next steps. e operation
sequence frequently appearing of the user in this scenario is
⟨1,liu,searchBtn,click,time1⟩,⟨2, liu,Inputbox,input,time2⟩,
⟨3,liu,OkBtn,click,time3⟩,⟨4,liu,Map,zoom in,time4⟩,
⟨5,liu,Map,zoom in,time5⟩,⟨6,liu,Map,click,time5⟩.e
system matched the rst three operations and then predicted
the next operation and adjusted the interface automatically.
e rst gure shows entering the search interface aer
clickingthesearchbutton,thesecondgureshowsthe
amplifying map automatically aer clicking the Ok button,
and the third gure shows the location information aer
clicking the site.
We reect the dynamic adaptive from two aspects mainly
from this experiment: the choice of transportation mode and
the user’s interest concerns.
6.2. User Evaluation and Analysis. We choose some test users
with certain discrimination and nish the appointed tasks.
We choose  testers by taking the dierences of users into
consideration in the user model. We choose half of the testers’
educationlevelsuchthatitisabovetheaverageandtheother
half is below the average. Including the testers, the prociency
can be divided into skilled, general, and strange. e sex ratio
is . :  and age distribution from  to  years old, which
were randomly selected.
e international organization for standardization (ISO)
includes the usability evaluation factors of a product which
xed tasks in a specic environment which are eectiveness,
interaction eciency, and user satisfaction. e eectiveness
is used to judge whether it can achieve certain functions
and interface supports the corresponding function. ere
are two functions of the testing interface: providing the
transportation recommended automatically and providing
the interest recommendation when the user nds route. e
user satisfaction is the subjective satisfaction of the user
interface. e evaluations of these two aspects are assessed
by the user survey. Interaction eciency is decided by error
rate, completion time, being easy to learn, and being easy to
use. We can record the error time and completion time of two
tasks and obtain testers’ evaluations about being memorable,
easy to learn, and easy to use and the eciency. We let
the testers to complete two contrast tasks under the same
condition and get some pairs of observe values. Analyze
these values to draw inferences. e dierence result which
is gotten from same testers in the same environment can
be regarded as the dierences made by dierent system. We

 Computational Intelligence and Neuroscience
F : Route searching.
F : POI searching.
F : Map operation.

Computational Intelligence and Neuroscience 
use general navigation system to do the comparison test and
obtain independent observations in pairs.
𝑋𝑖represents the time spent by comparison system to
complete a task, and 𝑌𝑖represents the time spent by test
system to complete a task. Suppose there are 𝑛pairs of
independent observations: (𝑋𝑖,𝑌
𝑖)(1≤ 𝑖≤𝑛).𝐷𝑖=𝑋
𝑖−
𝑌𝑖(1≤𝑖≤𝑛)is the dierence of 𝑋𝑖and 𝑌𝑖;observationsof
these values’ sample mean and sample variances are recorded
as 𝑑and 𝑠2
𝑑. e rejection region is |𝑡| = |𝑑/(𝑠𝑑/√𝑛)| ≥
𝑡𝑎/2(𝑛 − 1).
e rst task is to choose a vehicle and nd the route to
reach somewhere. e second task is nding interest point
around. And the third task is the map operation. At the
beginning of the experiment stage let the users be familiar
with the system and task. Record the time of each task
completedandthetimeoftotaltaskcompletedinthecourse
of the experiment. Count the total number of errors which
emerged in the processes of the tests using. And let each
test personnel complete the questionnaire at the end of the
experiment. e time of users, which are familiar with this
kind soware, to complete the rst task is usually about  s,
and the time of users to nd the corresponding route by
using the system which records the users’ selection and oers
suggestions is about  s. ese times are average values of
testers. For the rst task 𝑑≈6,𝑠𝑑≈ 40.714,𝑡0.05(49) =
1.6794,|𝑡| ≈ 1.032 which are out of the rejection region.
For the second task 𝑑≈7,𝑠𝑑≈ 26.589,𝑡0.05(49) = 1.6794,
|𝑡| = 1.8429 which fall into the rejection region and 𝑑≈5,
𝑠𝑑≈ 20.533,𝑡0.05(49) = 1.6794,|𝑡| = 1.7046 which fall into
the rejection. e fundamental task needs less average time,
but the advantage is not obvious through the calculation. In
the second task and the third task, operating time reduces
signicantly through the calculation. e error rate has little
dierence between the contrast system and the test system.
We can see that other factors are higher than contrast system
except memorability. In addition, eciency of the test system
improves obviously.
en we get the user’s satisfaction degree of the interface
on the aspects like being memorable, easy to learn, and easy
to use and eciency through the questionnaire survey. ese
 testers give the scores of two systems using experience and
get the average of each index, respectively, like Figure .We
can see that memorability of the traditional human-computer
interface is higher than the adaptive interface for elements
of adaptive human-computer interface are changeable. On
the other aspect, the scores of text system are higher than
the contrast system which eciency is greatly improved. To
illustrate the adaptation of the text system is greatly improved.
7. Conclusion
is paper proposes a context-aware adaptive human-com-
puter interface model for mobile LBS which is based on the
user model and described in three aspects: static composed
elements, dynamic interactive behavior, and adaptive strat-
egy. e adaptive user interface proposed in this paper has
advantages compared with traditional adaptive user interface
0
2
4
6
8
10
Memorable Easy to learn Easy to use Eciency
Test system
Contrast system
F : Questionnaire statistics.
as follows: () avoiding the limitation of the traditional
adaptive user interface caused by user classication and
achieving the adaptation according to the combination of
each user’s habits and external experiments; () paying more
attention to the dynamic interaction process and adjusting
the user interface in the interactive process more in line with
the real-time interaction; () using the context information
dynamically and then making the context information using
more eective.
e adaptive system based on the model in this paper
has some deciency; for example, the range of adaption
should be extended, and also there are limitations of the
current research to gain more eective information on user
knowledge,ability,andsoon.Inthefuture,wewillfurther
optimize the stimulus-judgment method, more eectively
use context information, enhance the adaptive result, and
improve the interface layout mechanism to reach the goal of
smoothandnaturalinterfaceadaptive.
Conflict of Interests
e authors declare that there is no conict of interests
regarding the publication of this paper.
Acknowledgments
e work is supported by the National Nature Science
Foundation of China (), the Science Foundation of
China for Postdoctors (no. M), the Natural Sci-
enceFoundationProjectofChongqing(cstcjcyjA),
and the Scientic and Technological Research Program of
Chongqing Municipal Education Commission (KJ).
References
[] X. Cao, “Designing mobile phone interface with children,” in
Proceedings of the Extended Abstracts on Human Factors in
Computing Systems (CHI EA ’07), pp. –, San Jose, Calif,
USA, April-May .
[] L. Zouhaier, Y. B. Hlaoui, and L. Jemni Ben Ayed, “Building
adaptive accessible context-aware for user interface tailored to
disable users,” in Proceedings of 37th IEEE Annual Computer
Soware and Applications Conference Workshops (COMPSACW
’13), pp. –, .

 Computational Intelligence and Neuroscience
[] E. Rich, “User modeling via stereotypes,” Cognitive Science,vol.
, no. , pp. –, .
[] A. Kobsa, “Generic user modeling systems,” User Modeling and
User-Adapted Interaction, vol. , no. -, pp. –, .
[] I.Poggi,C.Pelachaud,andB.deCarolis,“Todisplayornot
to display? Towards the architecture of a reexive agent,” in
Proceedings of the 2nd Workshop on Attitude, Personality and
Emotions in User-Adapted Interaction (UM ’01),.
[] Y.-T. Fan, D.-X. Teng, H.-Y. Yang, C.-X. Ma, G.-Z. Dai, and H.-
A. Wang, “An adaptive user interface model based on experience
awareness,” Jisuanji Xuebao/Chinese Journal of Computers,vol.
, no. , pp. –, .
[] X. Shi and F. Bian, “Research on AHCI specic human-
computer interface of GIS,” Geomatics and Information Science
of Wuhan University,vol.,no.,pp.–,.
[] N. Lathia, L. Capra, D. Magliocchetti et al., “Personaliz-
ing mobile travel information services,” Procedia—Social and
Behavioral Sciences,vol.,pp.–,.
[] B. N. Schilit and M. M. eimer, “Disseminating active map
information to mobile hosts,” IEEE Network,vol.,no.,pp.
–, .
[] P. J. Brown, J. D. Bovey, and X. Chen, “Context-aware applica-
tions: from the laboratory to the marketplace,” IEEE Personal
Communications,vol.,no.,pp.–,.
[] G. D. Abowd, A. K. Dey, P. J. Brown, N. Davies, M. Smith,
and P. Steggles, “Towards a better understanding of context and
context-awareness,” in Handheld and Ubiquitous Computing:
First International Symposium, HUC ’99 Karlsruhe, Germany,
September 27–29, 1999 Proceedings,vol.ofLecture Notes
in Computer Science, pp. –, Springer, Berlin, Germany,
.
[] W. Liu, X. Li, and D. Huang, “A survey on context awareness,” in
Proceedings of the International Conference on Computer Science
and Service System (CSSS ’11), pp. –, June .
[]D.Chu,N.D.Lane,T.T.-T.Laietal.,“Balancingenergy,
latency and accuracy for mobile sensor data classication,” in
Proceedings of the 9th ACM Conference on Embedded Networked
Sensor Systems (SenSys ’11),pp.–,Seattle,Wash,USA,
November .
[] M. Tomitsch, C. Ackad, O. Dawson, L. Hespanhol, and J.
Kay,“WhocaresabouttheContent?Ananalysisofplayful
behaviour at a public display,” in Proceedings of the International
SymposiumonPervasiveDisplays(PerDis’14),p.,.
[] J. Karat and C.-M. Karat, “e evolution of user-centered focus
in the human-computer interaction eld,” IBM Systems Journal,
vol. , no. , pp. –, .
[] G. D. Abowd, C. G. Atkeson, J. Hong, S. Long, R. Kooper, and
M. Pinkerton, “Cyberguide: a mobile context-aware tour guide,”
Wireless Network s,vol.,no.,pp.–,.
[] G. Cai, “Contextualization of geospatial database semantics for
human-GIS interaction,” GeoInformatica,vol.,no.,pp.–
, .
[] J. Chung and C. Schmandt, “Going my way: a user-aware route
planner,” in Proceedings of the SIGCHI Conference on Human
Factors in Computing Systems (CHI ’09), pp. –, ACM,
New York, NY, USA, April .
[] E. Ross, “Intelligent User Interfaces: Survey and Research Dire-
ctions,” , http://www.cs.bris.ac.uk.
[] D. Benyon, “Adaptive systems: a solution to usability problems,”
User Modelling and User-Adapted Interaction,vol.,no.,pp.
–, .
[] J.-W. Yoon, S.-I. Yang, and S.-B. Cho, “Adaptive mixture-of-
experts models for data glove interface with multiple users,”
Expert Systems with Applications, vol. , no. , pp. –,
.
[] S. Cheng and Y. Liu, “Eye-tracking based adaptive user inter-
face: implicit human-computer interaction for preference indi-
cation,” Journal on Multimodal User Interfaces,vol.,no.-,pp.
–, .
[] Z. Wang, N. Li, and B. Li, “Fast and independent access to
map directions for people who are blind,” Interacting with
Computers,vol.,no.,pp.–,.
[] S. Sulaiman and I. S. Sohaimi, “An investigation to obtain a
simple mobile phone interfacefor older adults,” in Proceedings of
the International Conference on Intelligentand Advanced Systems
(ICIAS ’10),pp.–,June.
[] W. A. E. F. Abdelsalam, Maintaining quality of service for adap-
tive mobile map clients [M.S. thesis],UniversityofWaterloo,
Ontario, Canada, .
[] A. Dillon and C. Watson, “User analysis in HCI - e historical
lessons from individual dierences research,” International
Journal of Human Computer Studies,vol.,no.,pp.–,
.
[] A. G. Christian and J. F. Ivonne, “A new behavioral measure
of cognitive exibility,” in EngineeringPsychologyandCognitive
Ergonomics. Understanding Human Cognition: 10th Interna-
tional Conference, EPCE 2013, Held as Part of HCI International
2013,LasVegas,NV,USA,July21–26,2013,Proceedings,Part
I,vol.ofLecture Notes in Computer Science,pp.–,
Springer, Berlin, Germany, .